{ "discovery_tags": [ "open-source", "playable", "free", "beginner-friendly", "starter-build", "became-startup", "used-by-millions", "weekend-build" ], "domains": { "accessibility": "Accessibility", "ai-ml": "AI & Machine Learning", "civic-tech": "Civic Tech", "climate": "Climate & Environment", "developer-tools": "Developer Tools", "education": "Education", "fintech": "Fintech", "games": "Games", "hardware": "Hardware", "healthcare": "Healthcare", "media-arts": "Media & Arts", "other": "Other", "robotics": "Robotics", "science": "Science", "social": "Social Impact", "web-mobile": "Web & Mobile" }, "projects": [ { "authors": [ "Sergej Kunz", "Anton Bukov" ], "body": "## What it does\n\n1inch is a decentralized exchange (DEX) aggregator. When you swap tokens, it splits the order across multiple liquidity sources to get a better average rate than any single exchange offers.\n\n## The key insight\n\nSergej Kunz and Anton Bukov built the first prototype in 56 hours at ETHGlobal New York in May 2019. The insight: on any single DEX, a large order moves the price against you (slippage). Splitting the same order across three DEXes with different liquidity pools produces a better average rate.\n\nThe math is simple. If exchange A holds 60% of the liquidity and exchange B holds 40%, sending 60% of your trade to A and 40% to B minimizes total slippage. The hard part is computing that split in real time across dozens of pools with changing reserves.\n\n## How it works\n\nThe system has three layers.\n\n**Routing.** Given a token pair and an amount, find all available pools (Uniswap, SushiSwap, Curve, and others). For each pool, compute the output amount given the input. Uniswap-style pools use a constant-product formula: x times y equals k.\n\n**Splitting.** Decide how to divide the input across pools. The original algorithm split the order iteratively, testing each split against the others. Later versions use on-chain optimization.\n\n**Execution.** Submit one transaction that performs all the sub-swaps atomically. If any leg fails, the whole transaction reverts. The user signs once; the contract handles the routing.\n\nThe aggregator takes a small fee (0.1% to 0.3%) from the price improvement it delivers. If it cannot beat the best single venue, it routes there instead.\n\n## Build something similar\n\nStart with two Uniswap V2-style pools for the same token pair, each with different reserves.\n\nWrite a function that, given an input amount, computes the output for pool A, pool B, and three split ratios (all A, 50/50, all B). Print the results.\n\nThe split that wins depends on the reserves. With equal reserves, 50/50 wins. With 80/20 reserves, the split roughly matches. This is the entire insight in ten lines of code.\n\nThen add a third pool. The number of split combinations grows, but the principle holds: more pools means more opportunities to beat any single venue.\n\n## Sources\n\nOrigin and prototype details are from the 1inch press room and blog posts linked in the evidence field.\n", "deep_dive": true, "description": "Decentralized exchange aggregator that finds the best token swap rates across DEX liquidity sources.", "domain": "fintech", "event": { "name": "ETHGlobal New York", "organizer": "ETHGlobal", "year": 2019 }, "evidence": [ { "supports": "History began at the ETHGlobal hackathon in New York, May 2019; 56-hour prototype", "url": "https://1inch.io/press-room" }, { "supports": "Built the first DEX aggregation protocol prototype at ETHGlobal NYC, May 19, 2019", "url": "https://1inch.com/blog/post/looking-back-on-5-years-of-innovation-1inch-in-2019-2" }, { "supports": "Founded May 2019 at the ETHGlobal hackathon in New York City", "url": "https://en.wikipedia.org/wiki/1inch" } ], "hot": 36, "id": "1inch", "key_idea": "Split one swap across multiple exchanges to get a better average price than any single venue offers.", "learn_from": "A hackathon prototype that solves one measurable problem (price slippage) can become infrastructure when the measurement matters to everyone.", "links": { "official": "https://1inch.io", "source": "https://github.com/1inch" }, "name": "1inch", "tags": [ "defi", "ethereum", "dex", "web3", "open-source" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\n2plant is a planning app for farmers. It combines soil moisture and evapotranspiration data to estimate when a field is at risk and when planting is more likely to succeed.\n\n## What makes it interesting\n\nIt won the Galactic Impact award at NASA Space Apps 2024. The entry takes two ordinary agricultural measurements and turns them into a timing decision, which is the decision a smallholder actually has to make.\n\n## What you can learn\n\nA satellite dataset becomes useful when it answers a question with a date attached. \"Is the soil wet enough this week\" is a better product than a soil-moisture map.\n\n## Limitations of this analysis\n\nThe record has no demo link and no field validation. Accuracy, coverage, and whether the app was used outside the competition are unknown.\n", "description": "App using soil moisture and evapotranspiration data to manage farmer planting risk.", "domain": "climate", "event": { "award": "Galactic Impact", "name": "NASA Space Apps Challenge", "year": 2024 }, "evidence": [ { "supports": "Space Apps 2024 Galactic Impact award", "url": "https://news.sciencex.in/nasa-international-space-apps-challenge-announces-2024-global-winners" } ], "hot": 25, "id": "2plant", "name": "2plant", "tags": [ "agriculture", "weather", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\n42 QuakeHeroes is a deep neural network that finds seismic events in time-series data. It reads waveform recordings and marks where an event begins.\n\n## What makes it interesting\n\nIt won Best Use of Technology at NASA Space Apps 2024. Detection is traditionally handled with hand-tuned thresholds and analyst review, so treating it as a learning problem changes what the system needs: labelled examples instead of tuned filters.\n\n## What you can learn\n\nA model is a way to encode the judgement of the people who labelled the data. The quality of the labels sets the ceiling, not the architecture.\n\n## Limitations of this analysis\n\nThe record does not describe the network, the training data, or the detection accuracy. It is an award listing, not an engineering report.\n", "description": "Deep neural network that locates seismic events in time-series data.", "domain": "science", "event": { "award": "Best Use of Technology", "name": "NASA Space Apps Challenge", "year": 2024 }, "evidence": [ { "supports": "Space Apps 2024 Best Use of Technology award", "url": "https://news.sciencex.in/nasa-international-space-apps-challenge-announces-2024-global-winners" } ], "hot": 25, "id": "42-quakeheroes", "name": "42 QuakeHeroes", "tags": [ "seismic", "ai-ml", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "authors": [ "Max Inferno" ], "body": "## What it is\n\nA Little to the Left is a tidy-up puzzle game. Each scene is a household mess, and the puzzle is working out the order the objects belong in.\n\n## The problem\n\nSorting games usually tell you the rule. This one asks you to infer it from the objects, which makes the mess the puzzle rather than the obstacle.\n\n## What makes it interesting\n\nIt started as a 48-hour entry for GMTK Game Jam 2020, on the theme \"Out of Control\". Annie and Lukas formed MAX INFERNO in 2021 and released the full game in November 2022. The jam build is still playable as a demo.\n\n## What you can learn\n\nA jam forces one scene to work before you build fifty. If the first tidy-up is satisfying, the content pipeline is the only thing left.\n\n## Limitations of this analysis\n\nThe record covers the origin, studio, and release. It does not describe the level design process or how the team scaled the jam concept.\n", "description": "Cozy tidying puzzle game about sorting household messes; began as a GMTK Game Jam 2020 entry and became a multi-platform commercial release.", "domain": "games", "event": { "name": "GMTK Game Jam 2020", "organizer": "Game Maker's Toolkit", "year": 2020 }, "evidence": [ { "supports": "Initially made for GMTK Game Jam 2020 (theme Out of Control, 48 hours) then expanded to a full game released November 8, 2022", "url": "https://en.wikipedia.org/wiki/A_Little_to_the_Left" }, { "supports": "Original jam submission of A Little to the Left (demo) for GMTK Game Jam 2020", "url": "https://itch.io/jam/gmtk-2020/rate/699181" }, { "supports": "Studio origin story: Annie and Lukas made their first game at GMTKJam 2020, formed MAX INFERNO in 2021", "url": "https://www.maxinferno.com/press" } ], "hot": 28, "id": "a-little-to-the-left", "links": { "demo": "https://maxinferno.itch.io/a-little-to-the-left", "official": "https://store.steampowered.com/app/1629520/A_Little_to_the_Left/" }, "name": "A Little to the Left", "tags": [ "gmtk-jam", "puzzle", "cozy", "commercial", "playable" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nA Normal Lost Phone is a narrative game played entirely through a found phone. You read messages, look at photos, and piece together who the owner was.\n\n## The problem\n\nStory games usually ask you to watch a character. A phone asks you to snoop, which is a more uncomfortable and more engaging position.\n\n## How the experience works\n\nThe phone is the interface. There is no map and no character to move. Progress comes from reading the apps and following the clues they contain.\n\n## What makes it interesting\n\nIt was built at Global Game Jam 2016 in Angoulême and later released on iOS, Android, and Steam. The interface was already familiar to every player, so the game needed no tutorial.\n\n## What you can learn\n\nUsing an interface people already know removes the cost of teaching it. The player spends their attention on the story instead of the controls.\n\n## Limitations of this analysis\n\nThe record does not cover the writing process, the fictional app design, or how the mobile and desktop versions differ.\n", "description": "Narrative game where players explore a found phone to uncover a personal story. Released on iOS, Android, and Steam.", "domain": "games", "event": { "name": "Global Game Jam 2016", "organizer": "Angoulême site" }, "evidence": [ { "supports": "GGJ 2016 Angoulême origin and platform releases", "url": "https://globalgamejam.org/ggj-success-stories" } ], "hot": 10, "id": "a-normal-lost-phone", "links": { "official": "https://lostphonegame.com/" }, "name": "A Normal Lost Phone", "tags": [ "game-jam", "narrative", "indie" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nAccount Kit was a Facebook login product that verified a phone number without sending an SMS. Early tests reported a 97% success rate.\n\n## The problem\n\nSMS verification fails for reasons the developer cannot see: delayed messages, wrong country codes, shared phones, and carrier filtering. Each failure is a user who does not get into the app.\n\n## What makes it interesting\n\nIt came out of the Facebook Seattle Platform team hackathon in 2016. The interesting choice is the constraint: prove the number without relying on a text message arriving.\n\n## What you can learn\n\nWhen a verification step has a high failure rate, the fix may be a different channel rather than a better error message. Measure the step, then remove its weakest dependency.\n\n## Limitations of this analysis\n\nThe record does not explain the verification mechanism, the markets it targeted, or how the 97% figure was measured.\n", "description": "Phone-number login verification without SMS. Achieved 97% success rate in early tests.", "domain": "developer-tools", "event": { "name": "Facebook Seattle Hackathon", "organizer": "Facebook", "year": 2016 }, "evidence": [ { "supports": "Facebook Seattle Platform team hackathon 2016", "url": "https://engineering.fb.com/2016/12/29/developer-tools/favorite-hacks-of-2016/" } ], "hot": 11, "id": "account-kit", "name": "Account Kit", "tags": [ "authentication", "facebook", "internal-hack-week" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nAdSense places contextual text and display ads on websites. Publishers paste a snippet of JavaScript; the script requests ads that match the page content and earns revenue when visitors click.\n\n## The key insight\n\nSusan Wojcicki proposed the idea in 2002 as a Google 20% project. The problem: Google had AdWords (advertisers bidding on keywords) but no way to place those ads on the millions of small websites that had content but no sales force. AdSense closed that gap by automating the match between page content and ad inventory.\n\nThe revenue number tells the story. Within two years of launch in June 2003, AdSense generated roughly 15% of Google's total revenue.\n\n## How it works\n\n**Content analysis.** The script sends the page URL (and sometimes extracted text) to Google's servers. An algorithm determines the page's topics: the words, the category, the context.\n\n**Ad selection.** Google matches those topics against keywords that advertisers are bidding on. Advertisers set bids per keyword in AdWords; AdSense surfaces the highest-bidding ads relevant to the page.\n\n**Auction.** Multiple ads compete for the slot. Google runs a second-price auction: the winner pays one cent more than the second-highest bidder. This keeps bids honest.\n\n**Rendering.** The script receives ad code (HTML and CSS) and injects it into the page. The publisher never sees the auction logic; they get a div that fills with ads.\n\n**Revenue share.** Google keeps a percentage (reported around 68%) and pays the publisher the rest (32%). The exact split is not public and has changed over time.\n\n## Build something similar\n\nBuild a page that displays a sponsored link based on the page's own content.\n\nStart with a static mapping: define five topics (cooking, finance, travel, fitness, technology) and one ad creative for each. Write a function that reads the page title and tags, scores each topic by word overlap, and picks the highest.\n\nThis gives you the core loop: analyze content, select an ad, render it. The difference between this and production AdSense is scale (millions of pages, real-time auction) and quality (advertisers actually paying).\n\nThen add a simple auction: two advertisers bidding on the same keyword, winner pays second-highest bid plus one cent. This is the AdWords mechanism.\n\n## Sources\n\nOrigin and revenue figures are documented in Wikipedia's 20% Project article, linked in the evidence field.\n", "deep_dive": true, "description": "Contextual advertising platform for publishers. Generated approximately 15% of Google revenue within two years of launch.", "domain": "web-mobile", "event": { "name": "Google 20% Time", "organizer": "Google", "year": 2003 }, "evidence": [ { "supports": "Google 20% project origin, Susan Wojcicki, revenue contribution", "url": "https://en.wikipedia.org/wiki/20%25_Project" } ], "hot": 20, "id": "adsense", "key_idea": "Match text ads to page content automatically, so publishers earn revenue without selling inventory themselves.", "learn_from": "A 20% time project won because it solved a real distribution problem: thousands of small sites had content but no sales team.", "links": { "official": "https://adsense.google.com/" }, "name": "AdSense", "notes": "Started by Susan Wojcicki as a 20% project, launched June 2003", "tags": [ "advertising", "google", "internal-hack-week" ], "verified_at": "2026-10-24" }, { "authors": [ "Soj Gamayon" ], "body": "## What it is\n\nAgriconnect is a credit marketplace for smallholder farmers in the Philippines. The prototype used a blockchain-based marketplace to connect farmers with lenders.\n\n## The problem\n\nSmall farms often lack the paperwork and collateral that formal lenders require. A marketplace can turn farming activity into a record that a lender can evaluate.\n\n## What makes it interesting\n\nIt won Red Bull Basement 2024 globally for Soj Gamayon of the Philippines. The pitch was access to credit, not a new crop or a new sensor.\n\n## What you can learn\n\nFor smallholder finance, the bottleneck is often evidence rather than money. A product that produces a trustworthy record can be more useful than a loan app.\n\n## Limitations of this analysis\n\nThe record documents the award and the concept only. There is no evidence of lending volume, default rates, or whether blockchain was necessary to the model.\n", "description": "Startup addressing agricultural credit access in the Philippines; won Red Bull Basement 2024 with a blockchain-based farming credit marketplace prototype.", "domain": "fintech", "event": { "award": "global winner", "name": "Red Bull Basement", "organizer": "Red Bull", "year": 2024 }, "evidence": [ { "supports": "Official winners list: Agriconnect - Soj Gamayon, Philippines 2024, for linking farmers to accessible credit", "url": "https://www.redbull.com/us-en/event-series/basement/red-bull-basement-past-winners" } ], "hot": 25, "id": "agriconnect", "links": { "official": "https://www.redbull.com/us-en/event-series/basement/red-bull-basement-past-winners" }, "name": "Agriconnect", "tags": [ "agritech", "red-bull-basement", "blockchain", "inclusion" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nAirocide is a line of air purifiers built on technology developed for spacecraft life support. The core process is catalytic oxidation, which breaks down organic gases rather than trapping them in a filter.\n\n## How the experience works\n\nAir passes through a catalyst bed, where volatile organic compounds are oxidized. The output is carbon dioxide and water vapor. Because the catalyst is not consumed, it does not need the replacement schedule of a particulate filter.\n\n## What makes it interesting\n\nThe product is a technology transfer: research paid for by a space program became a consumer appliance. NASA's spinoff programs exist to make that path visible and repeatable.\n\n## What you can learn\n\nDual-use research can produce a consumer product when the operating constraint is shared. A spacecraft and a sealed room both need air that does not accumulate trace gases.\n\n## Limitations of this analysis\n\nThe record does not document the license, the underlying studies, or independent performance testing. Claims about health effects are outside what the linked source establishes.\n", "description": "NASA-derived air purification technology commercialized after Space Apps-adjacent open innovation work; turned spacecraft-grade catalytic oxidation into consumer air cleaners.", "domain": "healthcare", "event": { "award": "technology transfer", "name": "NASA Innovation", "organizer": "NASA", "year": 2014 }, "evidence": [ { "supports": "NASA technology transfer programs commercialize spacecraft life-support research into consumer products including air purification", "url": "https://www.nasa.gov/directorates/space-tech/technology-transfer/" } ], "hot": 20, "id": "airocide", "links": { "official": "https://www.airocide.com" }, "name": "Airocide", "tags": [ "nasa", "air-purification", "commercialization", "health" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nAndroid Base Station turns a smartphone into a satellite ground station. The phone's radio receives open satellite data, and the app decodes it.\n\n## The problem\n\nGround stations are expensive equipment operated by institutions. A satellite that transmits open data still needs someone on the ground to listen.\n\n## What makes it interesting\n\nIt won Best Use of Hardware at NASA Space Apps 2014. The hardware was already in the participant's pocket; the project was mostly about proving the radio could do the job.\n\n## What you can learn\n\nBefore buying hardware, check what the device you own can already receive. A hackathon is a good place to test that boundary.\n\n## Limitations of this analysis\n\nThe record does not list supported satellites, the antenna setup, or the decoding chain. The project page is a competition entry, not a build guide.\n", "description": "Smartphone turned into a satellite ground station at NASA Space Apps 2014; won Best Use of Hardware for turning commodity phones into radio receivers for open satellite data.", "domain": "science", "event": { "award": "Best Use of Hardware", "name": "NASA Space Apps Challenge", "organizer": "NASA", "year": 2014 }, "evidence": [ { "supports": "Android Base Station project page for NASA Space Apps 2014; won Best Use of Hardware for phone-based satellite ground station", "url": "https://www.spaceappschallenge.org/2014/challenge/android-base-station/" } ], "hot": 20, "id": "android-base-station", "links": { "official": "https://www.spaceappschallenge.org/2014/challenge/android-base-station/" }, "name": "Android Base Station", "tags": [ "nasa-space-apps", "hardware", "radio", "satellite" ], "verified_at": "2026-10-24" }, { "authors": [ "Nick Farquhar" ], "body": "## What it is\n\nAppreciate is a property management platform that shares profits with property owners. It won Columbia Startup Weekend 2022 and later raised a $750K seed round.\n\n## The problem\n\nProperty managers usually charge a fee regardless of how the property performs. Owners carry the downside while the manager collects either way.\n\n## What makes it interesting\n\nThe model ties the manager's revenue to the owner's result. That is a business-model change, not a feature, and it is the kind of decision a weekend pitch can communicate clearly.\n\n## What you can learn\n\nA pricing model can be the product. If your incentive is visibly different from the incumbent's, the pitch explains itself.\n\n## Limitations of this analysis\n\nThe record does not cover the size of the portfolio, the terms of the profit share, or how the seed round was used.\n", "description": "Profit-sharing property management platform for property owners that won Columbia Startup Weekend 2022 and raised a $750K seed round.", "domain": "web-mobile", "event": { "award": "winner", "name": "Columbia Startup Weekend", "organizer": "Startup Weekend", "year": 2022 }, "evidence": [ { "supports": "Appreciate won Columbia Startup Weekend 2022; founder Nick Farquhar later closed a $750K seed round with EquipmentShare co-founder joining the executive team", "url": "https://startlandnews.com/2022/11/appreciate-columbia/" } ], "hot": 33, "id": "appreciate", "links": { "official": "https://www.appreciate.io/" }, "name": "Appreciate", "tags": [ "property-management", "startup-weekend", "proptech" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nArcobaleno turns astronomical images into sound. It maps two- and three-dimensional data to audio so that blind users can explore space imagery.\n\n## The problem\n\nAstronomy communication is almost entirely visual. A description of a nebula can tell you it is bright, but not where the brightness sits.\n\n## How the experience works\n\nImage properties such as position and intensity are mapped to sound parameters. Moving through the image changes what the listener hears, which gives spatial information without pixels.\n\n## What makes it interesting\n\nIt won the Global Connection Award at NASA Space Apps 2023. The project treats accessibility as a translation problem rather than a text alternative.\n\n## What you can learn\n\nSonification is not a fallback for images. It is a second channel that can carry information the visual version buries in color.\n\n## Limitations of this analysis\n\nThe record does not describe the sonification mapping or report testing with blind users. Both are essential to judging whether it works.\n", "description": "Sonifies 2D and 3D astronomical images so blind users can experience space imagery through sound.", "domain": "accessibility", "event": { "award": "Global Connection Award", "name": "NASA Space Apps Challenge", "year": 2023 }, "evidence": [ { "supports": "Space Apps 2023 Global Connection Award", "url": "https://www.spaceappschallenge.org/blog/the-nasa-space-apps-challenge-live-global-winners-announcement/" } ], "hot": 23, "id": "arcobaleno", "name": "Arcobaleno", "tags": [ "accessibility", "astronomy", "nasa-space-apps", "audio" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nAstrogenesis is an itinerary builder for planetary tourism. It assembles a trip concept from destination data and presents it as a plan.\n\n## What makes it interesting\n\nIt won Best Mission Concept at NASA Space Apps 2023. The entry reframes mission data as consumer planning, which is a different audience from the usual dashboard.\n\n## What you can learn\n\nThe same dataset can support a serious tool and a playful one. A tourism frame can teach a general audience where things are in the solar system.\n\n## Limitations of this analysis\n\nThe record does not include a demo link or describe the data sources. How the itineraries are generated is unknown.\n", "description": "Interactive planetary tourism itinerary builder.", "domain": "web-mobile", "event": { "award": "Best Mission Concept", "name": "NASA Space Apps Challenge", "year": 2023 }, "evidence": [ { "supports": "Space Apps 2023 Best Mission Concept award", "url": "https://www.spaceappschallenge.org/blog/the-nasa-space-apps-challenge-live-global-winners-announcement/" } ], "hot": 23, "id": "astrogenesis", "name": "Astrogenesis", "tags": [ "space", "web-mobile", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nAstroSweepers is a concept for removing orbital debris. It proposes a way to clear space junk from busy orbits.\n\n## The problem\n\nDefunct satellites and fragments stay in orbit for decades. Each piece is a collision risk for working spacecraft.\n\n## What makes it interesting\n\nIt won the Galactic Impact award at NASA Space Apps 2025. Debris removal is hard for reasons beyond engineering: someone has to pay for it, and no one owns the junk.\n\n## What you can learn\n\nA cleanup project needs a business or treaty answer before it needs a spacecraft. The hard part is often who authorizes the removal.\n\n## Limitations of this analysis\n\nThe record does not describe the capture mechanism, the target orbit, or any prototype. It is an award listing for a concept.\n", "description": "Orbital debris cleanup concept for removing space junk.", "domain": "science", "event": { "award": "Galactic Impact", "name": "NASA Space Apps Challenge", "year": 2025 }, "evidence": [ { "supports": "Space Apps 2025 Galactic Impact award", "url": "https://www.spaceappschallenge.org/2025/awards/global-winners/" } ], "hot": 25, "id": "astrosweepers", "name": "AstroSweepers", "tags": [ "space", "debris", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "authors": [ "Sophie Bolzer", "Nadine Szentivanyi" ], "body": "## What it is\n\nThe record for Audvice documents a trajectory rather than a product: a Red Bull Basement 2019 winner that later raised more than $3 million, worked with Fortune 500 companies, and was acquired.\n\n## What makes it interesting\n\nThe path is the point. A student competition entry became a funded company with enterprise customers and an exit, which is a much rarer outcome than a hackathon prize.\n\n## What you can learn\n\nA competition win is evidence that a team can pitch and ship under pressure. It is not evidence of a market, but it can open the door to the people who test one.\n\n## Limitations of this analysis\n\nThe record does not describe what Audvice did, its customers, or the acquisition terms. The product itself is not documented here.\n", "description": "Startup that won Red Bull Basement 2019; later raised more than $3 million in venture capital, worked with Fortune 500 companies, and was acquired.", "domain": "web-mobile", "event": { "award": "global winner", "name": "Red Bull Basement", "organizer": "Red Bull", "year": 2019 }, "evidence": [ { "supports": "Sophie Bolzer's idea won Red Bull Basement 2019; startup raised more than $3M VC, worked with Fortune 500 companies, and made a successful exit", "url": "https://www.redbull.com/us-en/red-bull-basement-sophie-bolzer-startup-success" }, { "supports": "Official past-winners page lists AUDVICE - Sophie Bolzer and Nadine Szentivanyi, Canada 2019", "url": "https://www.redbull.com/us-en/event-series/basement/red-bull-basement-past-winners" } ], "hot": 21, "id": "audvice", "links": { "official": "https://www.redbull.com/us-en/red-bull-basement-2019-winner-audvice-interview" }, "name": "Audvice", "tags": [ "red-bull-basement", "startup", "exit" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nBeFit is a social fitness app that rewards exercise with NFTs. It was described at the time as BeReal for pushups.\n\n## The problem\n\nFitness apps struggle with retention. A reward that has resale value is one way to make a streak worth protecting.\n\n## What makes it interesting\n\nIt won Best of ETHDenver and the Impact track at ETHDenver 2024. The design borrows a social format and attaches a crypto incentive, which is a common pattern at Ethereum hackathons and a hard one to keep honest.\n\n## What you can learn\n\nAn incentive only works if the reward costs less than the behavior it buys. Token rewards tend to attract the people who value the token most, not the people who need the habit.\n\n## Limitations of this analysis\n\nThe record does not report user numbers, retention, or token economics. There is no evidence the incentive worked.\n", "description": "Social fitness app that rewards exercise with NFTs. Described as BeReal for pushups.", "domain": "social", "event": { "award": "Best of ETHDenver and Impact track winner", "name": "ETHDenver", "year": 2024 }, "evidence": [ { "supports": "ETHDenver 2024 awards", "url": "https://unchainedcrypto.com/ethdenvers-2024-hackathon-onchain-security-ai-and-gaming-fun-dominate-the-winning-projects/" } ], "hot": 25, "id": "befit", "links": { "official": "https://befit.xyz/" }, "name": "BeFit", "tags": [ "web3", "fitness", "nft", "ethereum" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nBioJS is a JavaScript library for visualizing biological data in the browser: sequences, alignments, trees, and interaction networks.\n\n## The problem\n\nBioinformatics tools were often desktop applications or Python scripts. Sharing a result meant sending a file and asking the reader to install something.\n\n## How the experience works\n\nA developer imports a component, passes it data, and gets an interactive view in the page. Components are small and composable, so a paper's supplementary page can show the data directly.\n\n## What makes it interesting\n\nIn the 2014 Google Summer of Code, all five students stayed on as core developers and designed version 2.0. A summer program became the maintainer pipeline.\n\n## What you can learn\n\nOpen-source projects die when the original author leaves. Funding contributors, not just code, keeps a library alive.\n\n## Limitations of this analysis\n\nThe record does not cover the component architecture or the current maintenance status.\n", "description": "Interactive biological data visualization library. In 2014 Google Summer of Code, all five students became core developers and designed version 2.0.", "domain": "science", "event": { "name": "Google Summer of Code", "organizer": "Google", "year": 2014 }, "evidence": [ { "supports": "GSoC 2014, five students became core devs, v2.0 design", "url": "https://opensource.googleblog.com/2016/08/from-google-summer-of-code-to-game-of.html" } ], "hot": 15, "id": "biojs", "links": { "official": "https://biojs.io/", "source": "https://github.com/biojs/biojs" }, "name": "BioJS", "tags": [ "bioinformatics", "visualization", "open-source", "gsoc" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nCycles is a physically based render engine inside Blender. It traces light paths from the camera through the scene and produces images where reflections, shadows, and indirect light come from the same simulation instead of separate fakes.\n\n## The key insight\n\nRasterizers approximate light with shortcuts: shadow maps, reflection probes, ambient terms. Cycles does the opposite. It samples paths and averages them. Noise is not a bug, it is the visible cost of sampling, and it falls as the sample count rises.\n\nThat choice makes the engine predictable. A material that looks right in one scene tends to look right in another, because the same physical rules apply.\n\n## How it works\n\nA path tracer shoots rays from the camera. At each surface it chooses a bounce direction, weights the contribution, and continues. When a ray reaches a light or the background, the path contributes to the pixel. Many paths per pixel produce the estimate.\n\nImportance sampling decides where to send rays. Instead of sampling uniformly, the integrator samples directions that are likely to matter, such as toward a light or along a glossy lobe. Good sampling means fewer paths for the same noise level.\n\nAcceleration structures make ray traversal practical. A bounding volume hierarchy groups geometry so a ray can skip most of the scene. Without it, every bounce would test every triangle.\n\nMaterials are node graphs compiled into shader code. The graph describes the surface, and the renderer evaluates it at each hit. Because the description is procedural, the same material works in Cycles and, where supported, in the real-time engine.\n\nGPU rendering splits the same work across many cores. The integrator is data-parallel by design, which is why the same engine runs on CPU and GPU with different performance rather than different output.\n\n## Build something similar\n\nWrite a path tracer for spheres and a plane. Camera ray, intersect, bounce, accumulate. Render at one sample per pixel and look at the noise.\n\nAdd importance sampling for the light. The image should get cleaner without more samples.\n\nThen add one material property, such as roughness, and watch the highlight widen. The whole engine is an accumulation of these small physical decisions, which is why a minimal version teaches more than a feature-complete one.\n\n## Sources\n\nGSoC connection and project history are linked in the evidence field.\n", "deep_dive": true, "description": "Physically based render engine built into Blender. Creator Brecht Van Lommel worked on Blender via early Google Summer of Code.", "domain": "media-arts", "event": { "name": "Google Summer of Code", "organizer": "Google", "year": 2009 }, "evidence": [ { "supports": "GSoC connection, Brecht Van Lommel, 100+ Blender projects funded", "url": "https://www.blendernation.com/2019/09/30/see-the-neat-new-blender-features-from-google-summer-of-code-2019/" } ], "hot": 35, "id": "blender-cycles", "key_idea": "Simulate light transport instead of approximating it with shader tricks, and let samples trade time for accuracy.", "learn_from": "A physically based core can ship inside an existing tool if the integration is one engine behind an existing interface.", "links": { "official": "https://www.blender.org/features/rendering/", "source": "https://projects.blender.org/blender/blender" }, "name": "Blender Cycles", "tags": [ "rendering", "blender", "open-source", "gsoc" ], "verified_at": "2026-10-24" }, { "authors": [ "DANG!" ], "body": "## What it is\n\nBoomerang X is a first-person shooter built around one weapon: a boomerang that teleports you to wherever it lands.\n\n## How the experience works\n\nYou throw the boomerang, then recall it. Movement and combat use the same object, so positioning is the attack. The arena is designed around the arcs the boomerang can take.\n\n## What makes it interesting\n\nIt started as a 48-hour GMTK Game Jam entry and shipped through Devolver Digital in July 2021. A single verb carried both traversal and combat, which is why the jam version could demonstrate the whole game.\n\n## What you can learn\n\nIf one mechanic can serve two systems, the prototype already contains the design. The rest is content.\n\n## Limitations of this analysis\n\nThe record covers the jam origin and the release, not the level design or how the team tuned the boomerang.\n", "description": "Momentum-based first-person shooter built around a single teleporting boomerang; started as a 48-hour GMTK Game Jam entry and shipped via Devolver Digital in 2021.", "domain": "games", "event": { "name": "GMTK Game Jam", "organizer": "Game Maker's Toolkit" }, "evidence": [ { "supports": "itch page states the jam build was made in 48 hours for the GMTK game jam", "url": "https://dangstudios.itch.io/boomerang-x" }, { "supports": "Commercial release July 8, 2021, developed by DANG! and published by Devolver Digital", "url": "https://store.epicgames.com/p/boomerang-x" }, { "supports": "Follow-on studio news confirming Boomerang X as DANG!'s released title grown from the jam prototype", "url": "https://www.rockpapershotgun.com/boomerang-x-developer-dang-is-closing-after-failing-to-find-funding" } ], "hot": 25, "id": "boomerang-x", "links": { "demo": "https://dangstudios.itch.io/boomerang-x", "official": "https://store.steampowered.com/app/1256000/Boomerang_X/" }, "name": "Boomerang X", "tags": [ "gmtk-jam", "fps", "indie", "commercial", "playable" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nBroforce is a side-scrolling run-and-gun game built on action-movie archetypes. The roster is a set of parody heroes, each with a distinct weapon.\n\n## What makes it interesting\n\nIt grew out of a Ludum Dare 23 prototype called Rambros in 2012. The commercial game kept the destructible environments and the parody tone and added more heroes and levels.\n\n## What you can learn\n\nA jam prototype with a strong hook can scale through content rather than systems. If the first ten minutes work, the next hundred are production.\n\n## Limitations of this analysis\n\nThe record documents the prototype origin, not the level design, the engine, or the console ports.\n", "description": "Side-scrolling run-and-gun action game starring hyper-masculine action movie homages, born from a Ludum Dare prototype called Rambros.", "domain": "games", "event": { "name": "Ludum Dare 23", "year": 2012 }, "evidence": [ { "supports": "Broforce listed among games that originated from Ludum Dare submissions", "url": "https://en.wikipedia.org/wiki/Ludum_Dare" } ], "hot": 10, "id": "broforce", "links": { "official": "https://www.broforcegame.com" }, "name": "Broforce", "tags": [ "action", "side-scroller", "co-op" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nBuck is a build system for large codebases, built at Facebook for Android. It reads declarative build files, constructs a dependency graph, and executes independent work in parallel with caching.\n\n## The key insight\n\nThe prototype came out of a Facebook hackathon in 2012 because the Android build had become the slowest part of the developer loop. The fix was not a faster compiler. It was removing hidden dependencies so that work could be parallelized and reused.\n\nA build that recompiles the world on every change cannot be made fast by optimizing the compiler. It has to be restructured.\n\n## How it works\n\nBuild files declare targets: a library, a binary, a test, each with sources, dependencies, and rules. The declarations are explicit, so the tool can compute the full graph without scanning the filesystem for implicit inputs.\n\nFrom the graph, Buck derives the work. Independent nodes run concurrently. A node whose inputs and rule have not changed is skipped entirely, which is why incremental builds are fast even when the graph is large.\n\nCaching extends the skip across machines. If a target's inputs hash to a known value, the output can be fetched instead of rebuilt. This turns a build into a lookup for most nodes and a compile for the few that changed.\n\nCorrectness depends on the dependency declarations being complete. An undeclared dependency produces a build that works locally and fails in a clean checkout, so the tool enforces strictness rather than tolerating implicit inputs.\n\nFor Android specifically, the system also split application packaging from code compilation. Resources and code can be updated independently, which shortens the install cycle during development.\n\n## Build something similar\n\nWrite a build file format with three fields: inputs, outputs, and a command. Parse it into a graph.\n\nRun the graph in topological order, skipping any node whose inputs are unchanged. Time a rebuild after touching one file. That number is the whole point of the tool.\n\nThen add a content hash for each output and a cache directory keyed by hash. If a clean build can reuse entries from a previous one, you have the core of a modern build system.\n\n## Sources\n\nFacebook engineering writeup and hackathon origin are linked in the evidence field.\n", "deep_dive": true, "description": "Fast modular Android build system used at Facebook. Prototype built in one hackathon.", "domain": "developer-tools", "event": { "name": "Camp Hack-a-Thona", "organizer": "Facebook", "year": 2012 }, "evidence": [ { "supports": "Hack-a-Thona July 2012 prototype origin", "url": "https://engineering.fb.com/2013/05/14/android/buck-how-we-build-android-apps-at-facebook/" } ], "hot": 20, "id": "buck-build-system", "key_idea": "Describe a build as a graph of declared targets, then execute and cache each node independently.", "learn_from": "Find the bottleneck you can measure, then remove the mechanism that causes it. Build tools win on correctness and speed, not features.", "name": "Buck", "tags": [ "android", "build-tools", "facebook", "internal-hack-week" ], "verified_at": "2026-10-24" }, { "authors": [ "Leda Entertainment" ], "body": "## What it is\n\nBullion is a game concept that began at Global Game Jam 2016 and was later developed with Leda Entertainment. It was recognized at the Big Indie Pitch in 2022.\n\n## What makes it interesting\n\nThe origin is a collaboration. Coder Ben Pritchard met 3D artist Matthew Isteed at his first Global Game Jam, and the pair kept working together after the weekend.\n\n## What you can learn\n\nA game jam is a matchmaking event as much as a production sprint. The team you meet can matter more than the prototype you finish.\n\n## Limitations of this analysis\n\nThe record does not describe the game's mechanics, release status, or the Big Indie Pitch result beyond the recognition.\n", "description": "Project where a coder and a 3D artist met at Global Game Jam 2016 and later developed the concept with Leda Entertainment, recognized at Big Indie Pitch in 2022.", "domain": "games", "event": { "name": "Global Game Jam 2016", "organizer": "Global Game Jam", "year": 2016 }, "evidence": [ { "supports": "GGJ success stories: coder Ben Pritchard met 3D artist Matthew Isteed at his first Global Game Jam in 2016; Bullion later won recognition at Big Indie Pitch PocketGamer Connects London 2022", "url": "https://globalgamejam.org/ggj-success-stories" }, { "supports": "Additional GGJ coverage of Bullion - The Curse of the Cut-Throat Cattle by Leda Entertainment", "url": "https://globalgamejam.org/news/jam-fame-global-game-jam-projects-became-successful-published-games" } ], "hot": 11, "id": "bullion", "name": "Bullion - The Curse of the Cut-Throat Cattle", "tags": [ "global-game-jam", "commercial", "indie", "team-building" ], "verified_at": "2026-10-24" }, { "authors": [ "Giorgos Mendrinos" ], "body": "## What it is\n\nCaius is an AI mapping system that identifies road and trail networks from imagery. It turns unpaved and unmapped routes into data.\n\n## The problem\n\nRural roads and trails are often missing from official maps. That gap makes delivery, emergency response, and planning harder than they need to be.\n\n## What makes it interesting\n\nIt began as a CASSINI Hackathon concept, moved into ESA BIC Greece and Startab, and produced a consumer app in 2025. The competition was the start of an incubation path, not the end of the project.\n\n## What you can learn\n\nA hackathon entry can be the first milestone in a funding application. Treat the demo as evidence for the next step, not the finished product.\n\n## Limitations of this analysis\n\nThe record does not describe the model, its accuracy, or which regions are covered.\n", "description": "AI-based mapping system that automatically identifies road and trail networks; grew from a CASSINI Hackathon concept into a rural digitisation startup with ESA incubation and a 2025 consumer app.", "domain": "other", "event": { "name": "CASSINI Hackathon", "organizer": "European Union / ESA" }, "evidence": [ { "supports": "EU Agency for the Space Programme success story: Caius began as a CASSINI Hackathon idea, entered ESA BIC Greece and Startab, won regional mapping contracts, and launched consumer app Druiter in 2025", "url": "https://www.euspa.europa.eu/newsroom-events/success-stories/caius-hackathon-idea-rural-digitisation-startup" } ], "hot": 10, "id": "caius", "links": { "official": "https://www.euspa.europa.eu/newsroom-events/success-stories/caius-hackathon-idea-rural-digitisation-startup" }, "name": "Caius", "tags": [ "mapping", "ai", "space-data", "startup" ], "verified_at": "2026-10-24" }, { "authors": [ "Quek Siu Rui", "Lucas Ngoo", "Marcus Tan" ], "body": "## What it is\n\nCarousell is a mobile-first marketplace for secondhand goods. A seller photographs an item, adds a price, and publishes a listing in about a minute.\n\n## The problem\n\nClassifieds sites were built for desktop browsers and long forms. Selling a used phone required a camera, a cable, and patience, when the phone in the seller's pocket was already the best tool for the job.\n\n## How the experience works\n\nThe listing flow starts with a photo and asks for the minimum: title, price, category. Buyers browse and message the seller. The short flow is the product.\n\n## What makes it interesting\n\nThe team built the working demo in 54 hours at Startup Weekend Singapore in 2012, under the name SnapSell, and won first place. The name changed; the one-minute listing stayed.\n\n## What you can learn\n\nRemoving steps from a listing flow increases supply, and supply is what makes a marketplace usable. The camera was the feature, not the form.\n\n## Limitations of this analysis\n\nThe record covers the origin and the demo, not the trust and safety system or the payment model.\n", "description": "Mobile-first classifieds marketplace where selling an item takes about a minute with a photo.", "domain": "web-mobile", "event": { "award": "First place", "name": "Startup Weekend Singapore", "organizer": "Startup Weekend", "year": 2012 }, "evidence": [ { "supports": "Team SnapSell won the top prize at Startup Weekend (March 2012)", "url": "https://blog.carousell.com/carousell-pro-tips/featured-spotlight/carousell-early-years" }, { "supports": "Built a working demo in 54 hours at Startup Weekend Singapore 2012; overall winner", "url": "https://enterprise.nus.edu.sg/startup-story/carousell/" } ], "hot": 30, "id": "carousell", "links": { "official": "https://carousell.com" }, "name": "Carousell", "tags": [ "marketplace", "mobile", "secondhand" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nCelebrate Pride let Facebook users apply a rainbow filter to their profile picture. It shipped in 72 hours around a hackathon and reached roughly 26 million users in its first weekend.\n\n## The key insight\n\nThe feature is a filter, not a campaign site. It reused a surface every user already had, the profile photo, and changed one property of it. There was no new page to visit and no new account to create.\n\nThat is why it could spread in a weekend. The action was one tap, and the result was visible to everyone who looked at the profile.\n\n## How it works\n\nThe client offers a toggle on the existing profile picture flow. Choosing it sends the image reference to a service that composites the overlay and returns a new image. The original photo stays intact, so the change is reversible.\n\nThe image work is the bottleneck. Millions of users applying the same overlay within hours means the compositing step has to be cheap and the result has to be cacheable. The overlay is a fixed asset, so the output can be produced once per photo and served from cache rather than recomputed per view.\n\nClient support is the second constraint. The feature has to appear on every platform that shows a profile picture, which means the toggle and the result must be represented in the shared profile model rather than in one client's local state.\n\nTimeboxing shaped the architecture. A 72-hour window rules out new infrastructure, so the team built on existing image and profile services. The constraint produced a simpler design than a greenfield build would have.\n\n## Build something similar\n\nPick a surface every user already has, such as an avatar, a banner, or a display name. Add one reversible variant of it.\n\nPrecompute the variant as a static asset and serve it from cache. Then add the toggle to the shared user model so every client sees the same state.\n\nMeasure how long it takes a new user to reach the changed state. If it is more than a couple of taps, the feature is too expensive to spread.\n\n## Sources\n\nEngineering writeup, user count, and 72-hour timeline are linked in the evidence field.\n", "deep_dive": true, "description": "Rainbow profile picture filter. 26 million users in the first weekend. Intern hackathon project shipped globally in 72 hours.", "domain": "social", "event": { "name": "Facebook Hackathon", "organizer": "Facebook", "year": 2015 }, "evidence": [ { "supports": "Facebook hackathon origin, 26M users, 72-hour ship", "url": "https://engineering.fb.com/2015/07/02/android/72-hours-to-launch-celebrate-pride/" } ], "hot": 31, "id": "celebrate-pride", "key_idea": "Ship one gesture, a rainbow overlay on a profile photo, and let users supply the meaning.", "learn_from": "When the moment is the deadline, cut scope to a single action and make it work at full traffic.", "name": "Celebrate Pride", "tags": [ "facebook", "lgbtq", "internal-hack-week" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nChainlink for Swift is an integration between blockchain oracles and Swift's interbank messaging. It was selected as the winner of the Swift Hackathon 2025 Business Challenge.\n\n## The problem\n\nFinancial institutions already move value over Swift. Blockchains hold assets that those institutions want to settle, but the two systems do not share a message format.\n\n## What makes it interesting\n\nChainlink's co-founder presented the work at Sibos, the industry conference where interbank infrastructure gets discussed. A hackathon result reaching that stage is unusual.\n\n## What you can learn\n\nIntegrating with existing rails beats replacing them. The winning design usually meets the institution where it already operates.\n\n## Limitations of this analysis\n\nThe record does not describe the integration architecture, the assets involved, or whether any institution has adopted it.\n", "description": "Blockchain oracle integration selected as winner of the Swift Hackathon 2025 Business Challenge; Chainlink co-founder showcased the work at Sibos.", "domain": "fintech", "event": { "award": "Business Challenge winner", "name": "Swift Hackathon", "organizer": "Swift", "year": 2025 }, "evidence": [ { "supports": "Chainlink selected as winner of the Swift Hackathon 2025 Business Challenge; Sergey Nazarov showcased blockchain and oracle technologies at Sibos", "url": "https://chain.link/blog/chainlink-wins-swift-hackathon" } ], "hot": 35, "id": "chainlink-swift", "links": { "official": "https://chain.link/blog/chainlink-wins-swift-hackathon" }, "name": "Chainlink for Swift", "tags": [ "chainlink", "swift", "oracle", "banking", "hackathon-winner" ], "verified_at": "2026-10-24" }, { "authors": [ "Vasco Singh" ], "body": "## What it is\n\nChatEDU is a study assistant that turns course material into answers, practice sets, and flashcards. It accepts PDFs, slides, notes, and lecture video.\n\n## The problem\n\nStudents often have the material but not enough practice questions. Making them by hand is slow, and generic question banks do not match the course.\n\n## How the experience works\n\nA student uploads the material and asks questions. Answers cite the source passage, and the same content can be turned into practice problems or cards.\n\n## What makes it interesting\n\nFour Vanderbilt undergraduates built it for a Microsoft AI hackathon in 2023 because they lacked practice material for an algorithms final. It took first place for the Technical Project Prize.\n\n## What you can learn\n\nThe best hackathon project often solves the builder's own week. You already know the user, the workflow, and whether the result is good.\n\n## Limitations of this analysis\n\nThe record does not report answer accuracy or how the citations are verified. Both matter more than the feature list for a study tool.\n", "description": "Study assistant that turns PDFs, slides, notes and lecture video into citation-grounded chat, practice sets and flashcards; built by Vanderbilt students for a Microsoft AI hackathon.", "domain": "education", "event": { "award": "first place Technical Project Prize", "name": "Microsoft AI for Students Hackathon", "organizer": "Microsoft", "year": 2023 }, "evidence": [ { "supports": "Four Vanderbilt undergrads built ChatEDU for a Microsoft AI hackathon after lacking practice material for an algorithms final; later reached about $6,779 MRR and won five hackathons", "url": "https://www.sofarbot.com/stories/chatedu-vasco-singh-6704-mrr" }, { "supports": "Devpost summary of the student Azure AI education hackathon where ChatEDU took first place for the Technical Project Prize", "url": "https://info.devpost.com/customer-stories/ai-hackathons-on-devpost-2024" } ], "hot": 33, "id": "chat-edu", "links": { "official": "https://www.sofarbot.com/stories/chatedu-vasco-singh-6704-mrr" }, "name": "ChatEDU", "tags": [ "ai", "study-tools", "edtech", "microsoft-hackathon" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nConnected Earth Museum is a virtual museum about the climate, hosted by an AI guide. It presents climate data as 3D and 2D exhibits.\n\n## What makes it interesting\n\nIt won the Art and Technology award at NASA Space Apps 2024. The entry treats a dataset as an exhibition, with a host and a route through the material.\n\n## What you can learn\n\nA guided narrative can make scientific data approachable without simplifying it. The host does the work a curator would.\n\n## Limitations of this analysis\n\nThe record does not link a demo or describe the AI host. Whether the exhibits are accurate or the guide stays grounded is unknown.\n", "description": "AI-hosted immersive virtual climate museum with 3D and 2D visualizations.", "domain": "media-arts", "event": { "award": "Art and Technology", "name": "NASA Space Apps Challenge", "year": 2024 }, "evidence": [ { "supports": "Space Apps 2024 Art and Technology award", "url": "https://news.sciencex.in/nasa-international-space-apps-challenge-announces-2024-global-winners" } ], "hot": 25, "id": "connected-earth-museum", "name": "Connected Earth Museum", "tags": [ "climate", "ai-ml", "virtual-museum", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nDeadAss is a wearable that detects emergencies and acts on the wearer's behalf. It can secure funds and contact the right people when the wearer cannot.\n\n## The problem\n\nAn emergency response usually needs three things at once: detection, a decision, and money. A person in danger may be able to supply none of them.\n\n## What makes it interesting\n\nIt won the AI Hardware category at ETHGlobal Bangkok 2024. Combining a wearable with on-chain funds is unusual: the device has to act without a confirmation step.\n\n## What you can learn\n\nA system that acts for the user must define what happens when it is wrong. False positives are a design problem, not an edge case.\n\n## Limitations of this analysis\n\nThe record does not describe the sensors, the detection threshold, or any false-positive testing.\n", "description": "AI wearable that detects emergencies and secures funds and contacts when the wearer is in danger.", "domain": "hardware", "event": { "award": "AI Hardware winner", "name": "ETHGlobal Bangkok", "year": 2024 }, "evidence": [ { "supports": "ETHGlobal Bangkok 2024 AI Hardware winner", "url": "https://www.recursive.so/hackathons" } ], "hot": 25, "id": "deadass", "name": "DeadAss", "tags": [ "web3", "hardware", "ai-ml", "wearable" ], "verified_at": "2026-10-24" }, { "authors": [ "Team Gummy" ], "body": "## What it is\n\nDie in Dungeon is a roguelike built for the Spooktober game jam in 2022 and later expanded into a Steam release.\n\n## What makes it interesting\n\nThe jam prototype did the two things a store page needs: it looked like a finished game, and its demo attracted enough wishlists to justify the full build.\n\n## What you can learn\n\nA jam build can double as market research. Wishlist numbers tell you whether to spend a year on the idea before you spend it.\n\n## Limitations of this analysis\n\nThe record does not describe the combat, the progression, or how much of the jam code survived into the release.\n", "description": "Roguelike built for the Spooktober game jam; expanded into a full Steam release with a demo that drew large wishlist numbers.", "domain": "games", "event": { "award": "jam prototype", "name": "Spooktober Game Jam", "organizer": "Spooktober", "year": 2022 }, "evidence": [ { "supports": "Die in Dungeon Steam page credits a Spooktober jam origin and lists a full commercial release with demo", "url": "https://store.steampowered.com/app/1898250/Die_in_Dungeon/" } ], "hot": 33, "id": "die-in-dungeon", "links": { "official": "https://store.steampowered.com/app/1898250/Die_in_Dungeon/" }, "name": "Die in Dungeon", "tags": [ "game-jam", "roguelike", "indie", "steam" ], "verified_at": "2026-10-24" }, { "authors": [ "Matt Hall", "John Watkinson" ], "body": "## What it is\n\nDocracy was an open-source platform for legal documents, described at the time as GitHub for contracts. Users could publish, compare, and reuse agreements.\n\n## The problem\n\nStandard contracts were locked inside law firms and template sites. Small companies needed the same documents but could not afford the drafting.\n\n## What makes it interesting\n\nIt won a TechCrunch Disrupt hackathon in May 2011 and raised $650,000 from First Round Capital months later. The pitch was simple enough to land in a weekend.\n\n## What you can learn\n\nBorrowing a metaphor from a known tool is a fast way to explain a new one. \"GitHub for contracts\" does more work than a feature list.\n\n## Limitations of this analysis\n\nThe record does not cover the licensing model, the quality of the documents, or why the company wound down.\n", "description": "Open-source legal document platform billed as a GitHub for contracts; won a TechCrunch Disrupt hackathon and raised a $650K seed round.", "domain": "web-mobile", "event": { "award": "hackathon winner", "name": "TechCrunch Disrupt Hackathon", "organizer": "TechCrunch", "year": 2011 }, "evidence": [ { "supports": "Docracy named a TechCrunch Disrupt hackathon winner in May 2011 as a GitHub-style site for free legal documents", "url": "https://techcrunch.com/2011/05/22/hackathon-winner-docracy-is-a-github-for-legal-documents/" }, { "supports": "Disrupt hackathon winner Docracy raised $650,000 from First Round Capital months later", "url": "https://techcrunch.com/2011/11/03/docracy/" } ], "hot": 30, "id": "docracy", "links": { "official": "https://www.docracy.com" }, "name": "Docracy", "tags": [ "legaltech", "techcrunch-disrupt", "marketplace" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nDome Keeper is a mining and defense game. You dig for resources under a dome, then return to defend it when a wave arrives.\n\n## How the experience works\n\nThe loop alternates between two modes. Underground, you choose what to mine and how far to push. On the surface, you spend what you gathered on defenses before the next attack.\n\n## What makes it interesting\n\nIt started as Dome Romantik for Ludum Dare 48 in 2021, where it placed ninth overall. The commercial version kept the loop and added content and polish.\n\n## What you can learn\n\nThe jam version already contained the tension: greed versus preparation. If that choice is interesting in 48 hours, it will hold for a full game.\n\n## Limitations of this analysis\n\nThe record covers the jam placement and the release, not the upgrade design or the wave balancing.\n", "description": "Mining-and-defense roguelite where players dig for resources under a dome while defending against waves of enemies.", "domain": "games", "event": { "name": "Ludum Dare 48", "year": 2021 }, "evidence": [ { "supports": "Originally created as Dome Romantik for Ludum Dare 48 by Bippinbits; placed ninth overall", "url": "https://en.wikipedia.org/wiki/Dome_Keeper" }, { "supports": "Dome Keeper listed among games that originated from Ludum Dare submissions", "url": "https://en.wikipedia.org/wiki/Ludum_Dare" } ], "hot": 13, "id": "dome-keeper", "links": { "official": "https://domekeeper.com" }, "name": "Dome Keeper", "tags": [ "roguelite", "mining", "defense" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nDremel is a system for interactive ad-hoc analysis over very large datasets. It stores data in a columnar format and evaluates queries across many machines.\n\n## The problem\n\nAnalysts wanted answers in seconds from tables too large for one machine. Batch systems could do the work but not at the speed of a question.\n\n## What makes it interesting\n\nIt started as a Google 20% project in 2006 and became the foundation for BigQuery. The design choice that mattered was treating a query as a tree that can be pushed down to many servers, rather than moving the data to one.\n\n## What you can learn\n\nInteractive analysis is an architecture decision, not a UI one. If the query cannot be split, no interface will make it fast.\n\n## Limitations of this analysis\n\nThe record does not cover the paper's design details, the query language, or the path from internal system to Cloud product.\n", "description": "Interactive ad-hoc analysis over big data. Became the foundation for Google BigQuery.", "domain": "developer-tools", "event": { "name": "Google 20% Time", "organizer": "Google", "year": 2006 }, "evidence": [ { "supports": "Google 20% project origin and BigQuery connection", "url": "https://en.wikipedia.org/wiki/20%25_Project" } ], "hot": 0, "id": "dremel", "name": "Dremel", "notes": "Built by Andrey Gubarev as a 20% project", "tags": [ "bigdata", "google", "internal-hack-week" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nDroneStalker is a real-time drone tracking and prediction system. It estimates where a drone is and where it is heading.\n\n## What makes it interesting\n\nIt placed in the top five of 300 teams at Junction 2025 in Espoo, Finland, with 1,500 participants. Tracking a small moving object in real time is a perception and latency problem, not a dashboard problem.\n\n## What you can learn\n\nA prediction is only useful if it arrives before the object moves. Latency is part of the model's accuracy.\n\n## Limitations of this analysis\n\nThe record does not describe the sensors, the model, or the tracking range.\n", "description": "Real-time drone tracking and prediction system; winner at Junction 2025, Europe's largest hackathon, placing top 5 of 300 teams and 1,500 participants in Espoo, Finland.", "domain": "robotics", "event": { "award": "Winner (top 5)", "name": "Junction", "organizer": "Junction / AaltoES", "year": 2025 }, "evidence": [ { "supports": "DroneStalker won Junction 2025, placing Top 5 out of 300 teams and 1,500 participants at Europe's largest hackathon in Espoo, Finland; real-time drone tracking and prediction with YOLOv8 and Gemini", "url": "https://codya.dev/project/2511-junction" } ], "hot": 25, "id": "dronestalker", "links": { "official": "https://github.com/takara-ai/DroneStalker" }, "name": "DroneStalker", "tags": [ "junction", "drone", "computer-vision", "hackathon-winner" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nDropifi was a customer communication platform, and the first African company accepted into the 500 Startups accelerator.\n\n## What makes it interesting\n\nIt started at Startup Weekend Accra in 2011. The accelerator acceptance is the milestone the record preserves: a weekend project in Ghana reaching a Silicon Valley program.\n\n## What you can learn\n\nA local startup scene needs a visible path out. One accepted application gives the next team a reference point.\n\n## Limitations of this analysis\n\nThe record covers the origin and the accelerator, not the product. Its features, customers, and eventual outcome are not documented in the linked source.\n", "description": "Customer communication platform. First African company accepted into 500 Startups.", "domain": "web-mobile", "event": { "name": "Startup Weekend Accra", "organizer": "GEN-Ghana", "year": 2011 }, "evidence": [ { "supports": "Startup Weekend Accra 2011 origin and 500 Startups acceptance", "url": "https://www.gengh.org/impact-stories/launched-at-startup-weekend-dropifi-takes-on-silicon-valley" } ], "hot": 10, "id": "dropifi", "name": "Dropifi", "tags": [ "saas", "customer-service", "startup-weekend" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nEasy Taxi was a ride-hailing app for Latin America and other emerging markets. It connected passengers with taxi drivers rather than private cars.\n\n## The problem\n\nHailing a taxi on the street was slow, unpredictable, and unsafe in many cities. A dispatch app could make the wait visible.\n\n## What makes it interesting\n\nThe concept was first presented at Startup Weekend Rio 2011 and won the competition. The company then expanded across Latin America before the global ride-hailing platforms arrived.\n\n## What you can learn\n\nWorking with existing taxi fleets is a different business from creating a new fleet. It trades control for supply on day one.\n\n## Limitations of this analysis\n\nThe record covers the origin and expansion, not the dispatch system, the driver economics, or the company's eventual outcome.\n", "description": "Mobile ride-hailing app that connected passengers with taxi drivers across Latin America and other emerging markets.", "domain": "web-mobile", "event": { "name": "Startup Weekend Rio", "year": 2011 }, "evidence": [ { "supports": "Concept first presented during Startup Weekend RIO 2011; won the entire competition", "url": "https://en.wikipedia.org/wiki/Easy_Taxi" }, { "supports": "EasyTaxi born from Startup Weekend 2011 in Rio de Janeiro; won Startup Weekend Rio 2011 and subsequent competitions", "url": "https://www.techinasia.com/two-rare-hackathon-success-stories" }, { "supports": "BBC profile of Easy Taxi born out of frustration with hailing taxis in Rio", "url": "https://www.bbc.com/news/business-27991612" } ], "hot": 10, "id": "easy-taxi", "links": { "official": "https://en.wikipedia.org/wiki/Easy_Taxi" }, "name": "Easy Taxi", "tags": [ "mobility", "mobile", "marketplace" ], "verified_at": "2026-10-24" }, { "authors": [ "Joanna Power", "Paramveer Bhachu" ], "body": "## What it is\n\nEco Laundry is a water-saving washing machine concept from Joanna Power and Paramveer Bhachu.\n\n## What makes it interesting\n\nIt won Red Bull Basement 2020 for the UK and reached the global final. The team joined Imperial College London's Greenhouse accelerator through the Grantham Institute, and the project led to TEDx talks.\n\n## What you can learn\n\nA student competition can turn a concept into a team with access to labs and mentors. The accelerator, not the prize, is the useful part.\n\n## Limitations of this analysis\n\nThe record documents the award and the accelerator, not the engineering. Water savings, prototype status, and commercial availability are not established by the linked source.\n", "description": "Water-saving washing machine concept by Joanna Power and Paramveer Bhachu; won Red Bull Basement 2020 for the UK and went on to a hardware accelerator and TEDx talks.", "domain": "climate", "event": { "award": "UK winner / global finalist", "name": "Red Bull Basement", "organizer": "Red Bull", "year": 2020 }, "evidence": [ { "supports": "Joanna Power and Paramveer Bhachu won Red Bull Basement 2020 with a water-saving washing machine; joined Imperial College London's Greenhouse accelerator via the Grantham Institute", "url": "https://www.redbull.com/za-en/red-bull-basement-winners-catch-up" } ], "hot": 23, "id": "eco-laundry", "links": { "official": "https://www.redbull.com/za-en/red-bull-basement-winners-catch-up" }, "name": "Eco Laundry (Water-Saving Washing Machine)", "tags": [ "red-bull-basement", "hardware", "sustainability", "student-innovation" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nEco-Metropolis is a city-building game that uses real environmental data. Players grow a city while the simulation responds to actual conditions.\n\n## What makes it interesting\n\nIt won Most Inspirational at NASA Space Apps 2024. A game is a way to let someone experience a dataset as consequences rather than as a chart.\n\n## What you can learn\n\nIf your data describes a trade-off, let the user make the trade. A simulation teaches the constraint faster than an explainer.\n\n## Limitations of this analysis\n\nThe record does not link a build or describe which datasets feed the simulation.\n", "description": "City-building sustainability game that uses real environmental data.", "domain": "education", "event": { "award": "Most Inspirational", "name": "NASA Space Apps Challenge", "year": 2024 }, "evidence": [ { "supports": "Space Apps 2024 Most Inspirational award", "url": "https://news.sciencex.in/nasa-international-space-apps-challenge-announces-2024-global-winners" } ], "hot": 25, "id": "eco-metropolis", "name": "Eco-Metropolis", "tags": [ "gaming", "climate", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nEnergy Shock is a fast action game built at Global Game Jam 2018. It was named PC Gamer's best free game of the week and later entered Steam Early Access.\n\n## What makes it interesting\n\nThe jam build was good enough to be recommended as a free game, which gave the team an audience before the commercial version existed.\n\n## What you can learn\n\nA finished small build is a marketing asset. Releasing it free can do more for a project than holding it back until launch.\n\n## Limitations of this analysis\n\nThe record does not describe the mechanics or how the Early Access version changed the jam design.\n", "description": "Fast-paced action game. Named PC Gamer best free game of the week at the jam; later entered Steam Early Access.", "domain": "games", "event": { "name": "Global Game Jam 2018", "organizer": "China and CiGA Chengdu site" }, "evidence": [ { "supports": "GGJ 2018 origin, PC Gamer recognition, Steam release", "url": "https://globalgamejam.org/ggj-success-stories" } ], "hot": 10, "id": "energy-shock", "links": { "official": "https://store.steampowered.com/app/852980/Energy_Shock/" }, "name": "Energy Shock", "tags": [ "game-jam", "action", "indie" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nENSPin keeps files referenced by ENS names available on IPFS. It pins the content so the link does not break when the original host stops serving it.\n\n## The problem\n\nA name that resolves to content is only useful if the content still exists. Pinning is the difference between a permanent record and a dead link.\n\n## What makes it interesting\n\nIt won first place in the ENS Infrastructure category at ETHGlobal Taipei 2025. The project works on the unglamorous half of decentralization: persistence.\n\n## What you can learn\n\nNaming and storage are separate problems. A naming system does not guarantee that anyone keeps the file.\n\n## Limitations of this analysis\n\nThe record does not describe the pinning service, the cost model, or how it decides what to keep.\n", "description": "Automatically pins ENS-linked IPFS files for permanent persistence.", "domain": "developer-tools", "event": { "award": "1st place ENS Infrastructure", "name": "ETHGlobal Taipei", "year": 2025 }, "evidence": [ { "supports": "ETHGlobal Taipei 2025 first place ENS Infrastructure award", "url": "https://apu.edu.my/news/apubccs-enspin-clinches-top-prize-ethglobal-taipei-2025" } ], "hot": 25, "id": "enspin", "links": { "official": "https://enspin.dev/" }, "name": "ENSPin", "tags": [ "web3", "ens", "ipfs", "ethereum" ], "verified_at": "2026-10-24" }, { "authors": [ "Jabbok Schlacks", "Willy Schlacks" ], "body": "## What it is\n\nEquipmentShare is a construction equipment rental company with jobsite technology. Contractors rent machines and track them through the platform.\n\n## The problem\n\nEquipment rental is fragmented and local. A contractor with a short job needs a machine nearby, and the owner needs to know where it is.\n\n## What makes it interesting\n\nThe idea started as a winning pitch at Columbia Startup Weekend in 2014. The company later listed on Nasdaq, which is a long way from a weekend demo.\n\n## What you can learn\n\nA marketplace can start with a spreadsheet and a phone. The software matters when the fleet is large enough that manual tracking fails.\n\n## Limitations of this analysis\n\nThe record does not describe the telematics, the rental terms, or the path to public markets.\n", "description": "Construction equipment rental and jobsite technology company that began as a Columbia Startup Weekend winning idea and later listed on Nasdaq.", "domain": "other", "event": { "award": "winner", "name": "Columbia Startup Weekend", "organizer": "Startup Weekend", "year": 2014 }, "evidence": [ { "supports": "Company states the idea started as a winning entry at a local startup weekend in 2014 and grew into a top US equipment rental provider", "url": "https://www.equipmentshare.com/about" }, { "supports": "Report on how EquipmentShare's start traces to Columbia Startup Weekend in 2014", "url": "https://www.columbiatribune.com/story/news/local/2022/02/26/equipmentshares-success-thanks-columbia-startup-weekend/6828282001" }, { "supports": "2017 coverage of the Y Combinator alum building a construction machinery marketplace after the weekend pitch", "url": "https://venturebeat.com/technology/equipmentshare-raises-20-million-for-construction-equipment-marketplace" } ], "hot": 30, "id": "equipmentshare", "links": { "official": "https://www.equipmentshare.com" }, "name": "EquipmentShare", "tags": [ "construction", "marketplace", "startup-weekend", "ipo" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nEvoland is an action-adventure game where the game itself evolves. Each upgrade unlocks a new visual style or mechanic, moving from monochrome Game Boy graphics to 3D.\n\n## What makes it interesting\n\nIt won the \"Evolution\" theme at Ludum Dare 24 in 2012, chosen from more than 1,400 entries. The theme and the mechanic were the same thing, which is why the prototype read as a complete idea.\n\n## What you can learn\n\nWhen the theme can become the mechanic, the jam entry explains itself. The constraint is the design.\n\n## Limitations of this analysis\n\nThe record covers the jam win and the commercial release, not the engine or how the graphical stages were produced.\n", "description": "Action-adventure game where players unlock graphical and mechanical upgrades that evolve the game from monochrome Game Boy style to full 3D.", "domain": "games", "event": { "name": "Ludum Dare 24", "year": 2012 }, "evidence": [ { "supports": "Evoland began life as a Ludum Dare 24 entry by Nicolas Cannasse; won the jam theme 'Evolution' out of 1,400+ entries", "url": "https://www.hardcoregaming101.net/evoland" }, { "supports": "Evoland listed among games that originated from Ludum Dare submissions", "url": "https://en.wikipedia.org/wiki/Ludum_Dare" }, { "supports": "Original Ludum Dare game made available online as the full Shiro Games release approached", "url": "https://www.pcgamesn.com/indie/evoland-trailer-shows-lifespan-action-adventure-games" } ], "hot": 10, "id": "evoland", "links": { "official": "https://evoland.shirogames.com" }, "name": "Evoland", "tags": [ "action-adventure", "rpg", "meta" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nThe Like button is a one-click signal of approval attached to posts, photos, pages, and eventually other sites. It turned a private bookmark-style action into a public, countable metric.\n\n## The key insight\n\nJustin Rosenstein built a prototype called \"Props\" at a Facebook internal hackathon in 2007. It sat unused for roughly two years. When it launched in 2009, the important decision was not the icon. It was scope: one bit of feedback, applied to everything, with the count visible to everyone.\n\nA five-point rating would have carried more information and spread less. One bit is cheap to give, cheap to display, and comparable across objects of different kinds.\n\n## How it works\n\nOn Facebook, a Like is a row in a social graph with a user, an object, and a type. Counting is a read over that edge set, which is why the number can be shown on every post without a separate counter to keep in sync.\n\nOff the main site, the button ships as a small snippet that renders a cross-origin frame. The frame isolates the host page's CSS from Facebook's, and it carries the session needed to attribute the click. That isolation is why the same few lines of embed code work on a blog, a news site, and a product page.\n\nAggregation is the hard part. A popular page can produce far more writes than reads, and the count has to survive that ratio. Caching and precomputed totals absorb most of the load, with the exact value reconciled in the background.\n\nThe visual language mattered as much as the mechanism. A thumb, a count, and a consistent position trained users across unrelated sites. The button became recognizable before it became useful on any particular page.\n\n## Build something similar\n\nAdd one-bit feedback to something you already run. A row with a user id, an object id, and a timestamp is enough.\n\nShow the count next to the control, and make the control work on a second surface, such as an embed or a second client. The moment two surfaces share the same edge table, you have the core of the system.\n\nThen measure the ratio of writes to reads on a popular item. That number decides whether you need a cache, a queue, or nothing at all.\n\n## Sources\n\nOrigin and prototype details are linked in the evidence field.\n", "deep_dive": true, "description": "One-click social interaction feature that let users express approval of posts, photos, and pages across the web.", "domain": "social", "event": { "name": "Facebook internal hackathon", "year": 2007 }, "evidence": [ { "supports": "The Like button originated at a Facebook internal hackathon in summer 2007; Justin Rosenstein built the prototype codenamed 'Props'", "url": "https://www.fastcompany.com/90780140/the-full-story-of-the-like-button" } ], "hot": 30, "id": "facebook-like-button", "key_idea": "Reduce approval to one bit and one click, then make that click available on any page on the web.", "learn_from": "The most portable interaction is the smallest one. A single bit can cross every surface you own.", "links": { "official": "https://en.wikipedia.org/wiki/Like_button" }, "name": "Facebook Like button", "tags": [ "social-network", "engagement", "web" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nLive Faces was a live-streaming prototype built at a Facebook hackathon in 2012. It demonstrated broadcasting from the platform and is credited as an influence on Facebook Live.\n\n## What makes it interesting\n\nA hackathon demo can become a product direction years later. The prototype proved the interaction before the infrastructure and the business case existed.\n\n## What you can learn\n\nSome prototypes are arguments, not products. Their job is to make a decision possible, and they succeed when someone remembers them at the right moment.\n\n## Limitations of this analysis\n\nThe record does not describe the streaming stack, the audience, or how directly it influenced the shipped product.\n", "description": "Live-streaming demo prototype that helped inspire the creation of Facebook Live.", "domain": "media-arts", "event": { "name": "Facebook Hackathon", "organizer": "Facebook", "year": 2012 }, "evidence": [ { "supports": "2012 hackathon prototype connection to Facebook Live", "url": "https://www.fastcompany.com/40422805/the-story-behind-the-hackathon-project-that-helped-inspire-the-creation-of-facebook-live" } ], "hot": 10, "id": "facebook-live-precursor", "name": "Live Faces (Facebook Live precursor)", "notes": "Precursor prototype, not the final Facebook Live product", "tags": [ "streaming", "facebook", "internal-hack-week" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nFigma Slides is a presentation tool built on Figma's design canvas. Teams design slides with the same tools they use for interfaces, then present them in the same file.\n\n## The key insight\n\nThe team did not build a presentation editor from scratch. It built a slide layer on top of the editor that already existed: same canvas, same document, same multiplayer session, same component system.\n\nThat reuse set the schedule. The hackathon project at Figma's Maker Week in late 2023 became a shipped product in about two years, which is short for a new surface with its own editing model.\n\n## How it works\n\nA deck is a Figma document. Each slide is a frame, and the slide grid is a layout constraint over those frames rather than a separate data structure. Because slides are frames, every existing design tool works on them: auto layout, components, styles, images.\n\nEditing is shared state. Multiple cursors and selections already work because the multiplayer layer operates on the document, not on a screen. Adding a presenter view means adding a way to read the same document, not a new sync protocol.\n\nThe slide-specific parts are narrow: a grid that arranges frames in order, a presenter mode that advances through them, and transitions between them. Those are additions on top of the existing model rather than a parallel implementation.\n\nThe hard part is the constraint. A design canvas is free-form, and a slide deck is sequential. The grid has to enforce order and alignment without taking away the freedom that makes the canvas useful.\n\n## Build something similar\n\nTake an editor you already have, or a simple canvas with draggable rectangles. Add a grid that arranges frames in reading order.\n\nAdd a presenter mode that shows one frame at a time and advances on a key press. That is a slide deck.\n\nThen add one feature that only makes sense because the frames are editable, such as a shared component for a footer. The product becomes more than a slideshow at the point where the design tools carry over.\n\n## Sources\n\nHackathon origin and development timeline are linked in the evidence field.\n", "deep_dive": true, "description": "Presentation design tool built on Figma's collaborative canvas, letting teams create and present slide decks together in real time.", "domain": "web-mobile", "event": { "name": "Figma Maker Week", "year": 2023 }, "evidence": [ { "supports": "Figma Slides started as a hackathon project; initial development took about 6 months starting at a hackathon in late 2023", "url": "https://newsletter.pragmaticengineer.com/p/building-figma-slides-with-noah-finer" }, { "supports": "Figma Slides originated at Maker Week, Figma's internal hackathon", "url": "https://review.firstround.com/how-to-make-your-product-idea-go-viral-inside-your-company-lessons-from-figma-slides" } ], "hot": 23, "id": "figma-slides", "key_idea": "Reuse the canvas, the document model, and the multiplayer engine, then add only the slide layer.", "learn_from": "The fastest new product borrows an existing engine. Ship the delta, not a second platform.", "links": { "official": "https://www.figma.com/slides" }, "name": "Figma Slides", "tags": [ "design", "collaboration", "presentations" ], "verified_at": "2026-10-24" }, { "authors": [ "Alexa Andrzejewski" ], "body": "## What it is\n\nFoodspotting was a visual guide to dishes. Instead of reviewing restaurants, users photographed specific dishes and recommended them.\n\n## The problem\n\nRestaurant reviews describe a place. Diners often want to know what a particular dish looks like before ordering it.\n\n## What makes it interesting\n\nIt came out of the Startup Weekend circuit in 2009, founded by Alexa Andrzejewski. It was cited as a success story for the format and was acquired by OpenTable in 2015.\n\n## What you can learn\n\nNarrowing a review to one object makes the content comparable. A dish is easier to rate and photograph than a restaurant.\n\n## Limitations of this analysis\n\nThe record does not cover the recommendation system, the acquisition terms, or the app's technical stack.\n", "description": "Visual food discovery app that emerged from the Startup Weekend ecosystem; became a widely covered example of hackathon-to-startup success.", "domain": "web-mobile", "event": { "award": "project origin", "name": "Startup Weekend", "organizer": "Startup Weekend", "year": 2009 }, "evidence": [ { "supports": "Foodspotting cited as Startup Weekend success story; Alexa Andrzejewski launched the visual food guide after the event circuit", "url": "https://www.mississippi.edu/ume/centers/innovation/docs/Startup_Weekend_Case_Study_Foodspotting.pdf" }, { "supports": "Foodspotting visual food guide acquired by OpenTable in 2015", "url": "https://en.wikipedia.org/wiki/Foodspotting" } ], "hot": 30, "id": "foodspotting", "links": { "official": "https://en.wikipedia.org/wiki/Foodspotting" }, "name": "Foodspotting", "tags": [ "startup-weekend", "food", "mobile", "marketplace" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nFrame is a museum app that lets a visitor photograph an artwork and get information about it. It was a co-winner at the Walters Art Museum's first Art Bytes hackathon in 2012.\n\n## The problem\n\nMuseum labels are fixed to walls. A visitor who wants context for a piece has to find the right plaque or wait for a tour.\n\n## How the experience works\n\nThe visitor takes a photo. The app matches the artwork and returns information about it, using the phone as the label.\n\n## What makes it interesting\n\nIt shared the top prize with Tanzaku, another entry from the same event. Both projects used the phone as the interface to the collection rather than building a separate guide.\n\n## What you can learn\n\nA museum already has the content. The hack is the lookup, not the encyclopedia.\n\n## Limitations of this analysis\n\nThe record does not describe the matching method or whether the app was deployed at the museum.\n", "description": "Smartphone app that frames museum artworks by photo; co-winner of the first Walters Art Museum Art Bytes hackathon in 2012 with Tanzaku, taking a $1,000 prize.", "domain": "media-arts", "event": { "award": "co-winner ($1,000)", "name": "Art Bytes", "organizer": "Walters Art Museum", "year": 2012 }, "evidence": [ { "supports": "Frame and Tanzaku each took top $1,000 prizes at the first Walters Art Museum Art Bytes hackathon in 2012; Frame used phone photos to frame artworks", "url": "https://technical.ly/software-development/art-bytes-these-are-the-eight-projects-from-the-first-ever-walters-art-museum-hackathon" } ], "hot": 20, "id": "frame-walters", "links": { "official": "https://technical.ly/software-development/art-bytes-these-are-the-eight-projects-from-the-first-ever-walters-art-museum-hackathon" }, "name": "Frame (Walters Art Museum)", "tags": [ "museum-hackathon", "mobile", "art", "walters" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nFriday Night Funkin' is a rhythm game where you press arrow keys in time with music while a opponent sings against you. It started as a Ludum Dare 47 entry in October 2020 and became a viral hit with a large modding scene.\n\n## The key insight\n\nRhythm games have a small mechanical core: notes arrive on a timeline, the player hits them within a timing window, and the game scores the difference. The jam build implemented that core and wrapped it in characters and music.\n\nThe team open-sourced the game. The chart format and the asset layout were readable, so modders could add songs and characters without reverse-engineering the engine. Most of the game's long tail came from that decision.\n\n## How it works\n\nAudio position is the clock. The game tracks the current playback time of the song and spawns notes whose timestamps fall inside the visible window. Judgement compares the time of the key press against the note's timestamp and assigns a rating from the offset.\n\nA chart is a sequence of notes with timestamps, lanes, and lengths for holds. Splitting the chart from the song file means a new track is data, not code.\n\nThe engine is written in Haxe and targets multiple platforms through OpenFL. That choice gives one codebase for desktop and web, which matters when the audience arrives through a browser game portal.\n\nModding works because assets and charts are loaded from a known directory structure. A mod is a folder with a song, a chart, and sprites. The game reads the folder and plays it, with no plugin API required.\n\n## Build something similar\n\nBuild the clock first. Load an audio file, display its playback time, and draw a marker that moves with it.\n\nAdd notes with timestamps. Spawn each note a fixed number of seconds before its time and move it toward a hit line. On a key press, compare the current time to the nearest note's timestamp and print the offset.\n\nThen move the notes into a JSON file. When you can add a song by editing data instead of code, you have the same extension point that made modding possible.\n\n## Sources\n\nLudum Dare origin, repository, and developer devlog are linked in the evidence field.\n", "deep_dive": true, "description": "Open-source rhythm game where a boyfriends battles ex-rockstar dads in rap-battle-style musical face-offs.", "domain": "games", "event": { "name": "Ludum Dare 47", "year": 2020 }, "evidence": [ { "supports": "Initially created for the Ludum Dare 47 game jam on October 5, 2020", "url": "https://en.wikipedia.org/wiki/Friday_Night_Funkin%27" }, { "supports": "Official repo states the game was originally made for Ludum Dare 47 'Stuck In a Loop'", "url": "https://github.com/FunkinCrew/Funkin/blob/v0.2.7.1/README.md" }, { "supports": "Developer ninjamuffin99's devlog describing the Ludum Dare 47 release and its rapid viral spread", "url": "https://ninja-muffin24.itch.io/friday-night-funkin/devlog/184825/friday-night-funkin-going-forward" } ], "hot": 48, "id": "friday-night-funkin", "key_idea": "A rhythm game is a chart plus a clock. Everything else is presentation, and presentation can be replaced.", "learn_from": "Keep the core loop tiny and the content format open, and a community will build the rest.", "links": { "official": "https://www.newgrounds.com/portal/view/770371", "source": "https://github.com/ninjamuffin99/Funkin" }, "name": "Friday Night Funkin'", "tags": [ "rhythm", "open-source", "viral" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\ng0v is a Taiwanese civic-tech community that builds open tools for government transparency: budget visualizations, legislative trackers, referendum monitors, and epidemic dashboards.\n\n## The problem\n\nGovernment data is published, but not in a form most people can read. Budgets arrive as documents; legislative activity arrives as records without a view.\n\n## How the experience works\n\nVolunteers take public data, clean it, and publish it as a browsable tool. Each project is independent, and the community meets at recurring hackathons rather than through a single organization.\n\n## What makes it interesting\n\nIt started with a late-2012 hackathon that produced a better national budget visualization. The portfolio grew to more than 40 hackathons, with tools for legislative monitoring, referendums, and COVID-19 data.\n\n## What you can learn\n\nCivic tech spreads through reusable artifacts. A budget parser or a data format lasts longer than a campaign, because the next team starts from it.\n\n## Limitations of this analysis\n\nThe record does not cover funding, project governance, or how volunteers maintain tools after the hackathon.\n", "description": "Taiwan civic-tech movement that builds tools for government transparency, including crowd-sourced dashboards and legislative trackers.", "domain": "civic-tech", "event": { "name": "g0v hackathon", "year": 2012 }, "evidence": [ { "supports": "g0v's own portfolio describing the movement's origins in late-2012 hackathons and its growth to 40+ hackathons", "url": "https://g0v.tw/intl/en/portfolio" } ], "hot": 10, "id": "g0v", "links": { "official": "https://g0v.tw" }, "name": "g0v (g0v)", "tags": [ "transparency", "taiwan" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nGalaxiga is a retro space shooter built at a Global Game Jam and later released on mobile and Steam.\n\n## What makes it interesting\n\nThe jam version established the arcade feel: fixed screen, waves of enemies, a score worth chasing. The commercial release added platforms and polish without changing the core.\n\n## What you can learn\n\nA genre with a known loop is a good jam choice. The audience already understands the controls, so the prototype can spend its 48 hours on feel.\n\n## Limitations of this analysis\n\nThe record does not describe the weapon system, the level structure, or the porting work.\n", "description": "Space shooter built during a Global Game Jam; expanded into a polished mobile and Steam release with retro arcade aesthetics.", "domain": "games", "event": { "award": "jam prototype", "name": "Global Game Jam", "organizer": "Global Game Jam", "year": 2018 }, "evidence": [ { "supports": "Galaxiga Steam page credits a Global Game Jam origin for the retro space shooter", "url": "https://store.steampowered.com/app/875240/Galaxiga/" } ], "hot": 31, "id": "galaxiga", "links": { "official": "https://store.steampowered.com/app/875240/Galaxiga/" }, "name": "Galaxiga", "tags": [ "game-jam", "arcade", "space", "mobile" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nGalcon is a real-time strategy game played with one click. You select a planet, then send a fleet to another; numbers decide the outcome.\n\n## The problem\n\nReal-time strategy usually means build orders and tech trees. Galcon removes both, leaving only the decision of where to send ships and when.\n\n## What makes it interesting\n\nIt entered Ludum Dare 8 in 2006 and won four first places. The design fits on one screen and one input, which is why it later worked on phones.\n\n## What you can learn\n\nRemoving a layer from a genre can create a new one. The strategy survives without the economy.\n\n## Limitations of this analysis\n\nThe record covers the jam result and the 2006 release, not the AI or the multiplayer netcode.\n", "description": "Real-time strategy game where players flood planets with fleets using a simple one-click interface.", "domain": "games", "event": { "name": "Ludum Dare 8", "year": 2006 }, "evidence": [ { "supports": "Galcon was an entry to the Ludum Dare 48-hour competition in 2006; announced winner of the 8th Ludum Dare with four 1st places", "url": "https://www.galcon.com/history" }, { "supports": "Galcon Classic released for desktop in late 2006 for Ludum Dare, then developed into a more advanced video game", "url": "https://en.wikipedia.org/wiki/Galcon" } ], "hot": 10, "id": "galcon", "links": { "official": "https://www.galcon.com" }, "name": "Galcon", "tags": [ "strategy", "multiplayer", "real-time" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nGarden Buddies is a cozy game about tending a garden with a companion. It was made during the May Day Mental Health game jam in 2022, run by Global Game Jam with DeepWell DTx.\n\n## What makes it interesting\n\nIt was featured at Wholesome Direct 2023 and shipped on Steam in October 2023. The jam had a mental-health brief, which shaped the pace and the subject matter.\n\n## What you can learn\n\nA themed jam can produce a game that would be hard to pitch otherwise. The constraint gives the team permission to make something quiet.\n\n## Limitations of this analysis\n\nThe record does not describe the mechanics or how the mental-health brief was handled in design.\n", "description": "Cozy game made during the May Day Mental Health game jam run by Global Game Jam with DeepWell DTx in 2022; featured at Wholesome Direct 2023 and shipped on Steam in October 2023.", "domain": "games", "event": { "name": "May Day Mental Health Game Jam", "organizer": "Global Game Jam with DeepWell DTx", "year": 2022 }, "evidence": [ { "supports": "GGJ news: Garden Buddies created during the 2022 May Day Mental Health jam with DeepWell DTx, featured at Wholesome Direct 2023, published on Steam October 2023", "url": "https://globalgamejam.org/news/jam-fame-global-game-jam-projects-became-successful-published-games" } ], "hot": 13, "id": "garden-buddies", "links": { "official": "https://store.steampowered.com" }, "name": "Garden Buddies", "tags": [ "global-game-jam", "mental-health", "wholesome", "indie" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nGephi is an open-source tool for exploring and visualizing graphs. You load a network, run a layout, color and size nodes by a metric, and filter the result interactively.\n\n## The key insight\n\nMost graph tools at the time produced static output: compute a layout, render an image, publish it. Gephi made layout interactive. You can start a force-directed layout, watch it settle, stop it, and change a parameter while looking at the result.\n\nThat changes the work. Exploration becomes a loop of adjusting and observing rather than a batch job with a preview at the end.\n\n## How it works\n\nThe data model is a graph with nodes, edges, and attributes, plus a separate attribute table. Metrics such as degree, betweenness, and modularity are computed into columns that can drive visual properties.\n\nLayout algorithms are iterative. Force-directed layouts treat edges as springs and nodes as repelling charges, then repeatedly move each node by the net force. The result is a position for every node, refined over many steps rather than solved once.\n\nRendering is the constraint. A graph with hundreds of thousands of edges cannot be drawn naively at interactive rates, so the view culls, batches, and uses the GPU for the heavy drawing. The layout runs in the background while the view stays responsive.\n\nImporters are the adoption path. Gephi reads edge lists, matrix formats, and database queries, because a visualization tool is only as useful as the data it can open. Google Summer of Code support from 2009 to 2011 funded much of that connector and plugin work.\n\n## Build something similar\n\nLoad a small edge list and compute a force-directed layout yourself. Springs for edges, repulsion for all pairs, and a damping factor. Run a hundred iterations and draw the result.\n\nThen make it interactive: a slider for repulsion, a button to stop and restart. The tool becomes useful when you can see the effect of a parameter without re-running a pipeline.\n\nAdd one metric, such as degree, and map it to node size. That single mapping turns a picture of a graph into an explanation of it.\n\n## Sources\n\nGSoC history and project description are linked in the evidence field.\n", "deep_dive": true, "description": "Open-source network analysis and visualization tool. Described as Photoshop for network data. Google Summer of Code from 2009 to 2011 accelerated its growth.", "domain": "science", "event": { "name": "Google Summer of Code", "organizer": "Google", "year": 2009 }, "evidence": [ { "supports": "GSoC 2009-2011 participation and community growth", "url": "https://opensource.googleblog.com/2012/03/gephi-how-3-years-of-google-summer-of.html" } ], "hot": 35, "id": "gephi", "key_idea": "Treat layout as the first operation on network data, not the last step of a report.", "learn_from": "A visualization tool lives on its layout algorithms and its ability to open other people's data formats.", "links": { "official": "https://gephi.org/", "source": "https://github.com/gephi/gephi" }, "name": "Gephi", "tags": [ "visualization", "open-source", "graph", "gsoc" ], "verified_at": "2026-10-24" }, { "authors": [ "Sam Zaid", "Jessica Scorpio", "Elliot Kroo" ], "body": "## What it is\n\nGetaround is a peer-to-peer car rental marketplace. Owners list cars by the hour or day; renters book and unlock them through the app.\n\n## The problem\n\nCar sharing usually requires a company fleet and a physical key handoff. Peer-to-peer rental has a trust problem and an access problem: the owner is not there to hand over the key.\n\n## How the experience works\n\nA hardware kit installed in the car connects the vehicle to the platform. After a booking, the renter unlocks the car from the phone, so the owner never has to meet them.\n\n## What makes it interesting\n\nThe company launched at TechCrunch Disrupt NYC 2011 and won the Startup Battlefield that day. It later listed on the NYSE, then announced a wind-down in 2026. That arc is unusually complete for an archive entry: weekend stage, public company, ending.\n\n## What you can learn\n\nMarketplaces with a physical access problem need hardware, not just software. The install and recovery path for that hardware is part of the product, not an accessory to it.\n\n## Limitations of this analysis\n\nThe record does not cover the kit's design, the insurance model, or the reasons behind the 2026 wind-down.\n", "description": "Peer-to-peer car-sharing marketplace with a hardware car-kit for phone unlock; won TechCrunch Disrupt NYC 2011, later NYSE-listed, wound down in 2026.", "domain": "other", "event": { "award": "winner", "name": "TechCrunch Disrupt NYC Startup Battlefield", "organizer": "TechCrunch", "year": 2011 }, "evidence": [ { "supports": "Getaround named winner of TechCrunch Disrupt NYC 2011 Startup Battlefield", "url": "https://techcrunch.com/2011/05/25/and-the-winner-of-techcrunch-disrupt-nyc-is-getaround" }, { "supports": "Company account of public launch at Disrupt NYC 2011 and eventual NYSE listing", "url": "https://getaround.com/blog/our-journey-from-small-startup-to-publicly-traded-on-the-nyse" }, { "supports": "Founded around the Disrupt era, won Disrupt NY 2011, later public; board announced dissolution and liquidation in June 2026", "url": "https://en.wikipedia.org/wiki/Getaround" } ], "hot": 20, "id": "getaround", "links": { "official": "https://www.getaround.com" }, "name": "Getaround", "tags": [ "car-sharing", "sharing-economy", "techcrunch-disrupt", "ipo" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nGods Will Be Watching is a point-and-click thriller. A stranded crew faces impossible choices with too few resources.\n\n## The problem\n\nSurvival games often let you optimize your way out. This one makes every option cost someone, which turns resource management into an ethical decision.\n\n## What makes it interesting\n\nIt started as a Ludum Dare 26 entry in 2013, made in 72 hours, and was later expanded into a full game. The original jam version is still playable.\n\n## What you can learn\n\nA prototype that makes the player uncomfortable is doing something right. The discomfort is the hook, and it survives the production.\n\n## Limitations of this analysis\n\nThe record does not describe the scenario design or how the expanded version changed the jam's choices.\n", "description": "Point-and-click thriller where a stranded crew faces impossible ethical choices with limited resources.", "domain": "games", "event": { "name": "Ludum Dare 26", "year": 2013 }, "evidence": [ { "supports": "Started as a Ludum Dare 26 game; original jam project still playable", "url": "https://www.engadget.com/2014-02-26-point-and-click-thriller-gods-will-be-watching-stares-down-a-jun.html" }, { "supports": "Made in 72 hours as part of the Ludum Dare game-making competition; developers expanded it into a fuller experience", "url": "https://www.theverge.com/2013/7/22/4545296/despair-is-a-game-in-gods-will-be-watching" } ], "hot": 10, "id": "gods-will-be-watching", "links": { "official": "https://www.deconstructeam.com/games/gods-will-be-watching/" }, "name": "Gods Will Be Watching", "tags": [ "narrative", "point-and-click", "moral-choice" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nGoogle News is an aggregator that groups articles by story and links readers to the publishers.\n\n## The problem\n\nFollowing a news event meant visiting many sites and comparing accounts. There was no index of who had covered what.\n\n## How the experience works\n\nThe service clusters articles about the same event, ranks the cluster, and presents the sources together. The reader chooses a story, then a publisher.\n\n## What makes it interesting\n\nIt grew out of a hobby project by Krishna Bharat and entered beta in September 2002, under Google's 20% time. The design has no editors: grouping and ranking are algorithmic, which is both the scale advantage and the recurring criticism.\n\n## What you can learn\n\nAutomation can cover a breadth no newsroom can. It also inherits every judgement the algorithm makes, so the ranking rule is an editorial policy whether or not anyone calls it one.\n\n## Limitations of this analysis\n\nThe record covers origin and launch, not the ranking system or publisher agreements.\n", "description": "News aggregator built from Krishna Bharat's hobby project. Entered beta September 2002.", "domain": "web-mobile", "event": { "name": "Google 20% Time", "organizer": "Google", "year": 2002 }, "evidence": [ { "supports": "Google 20% project origin, Krishna Bharat, 2002 beta", "url": "https://en.wikipedia.org/wiki/20%25_Project" } ], "hot": 0, "id": "google-news", "links": { "official": "https://news.google.com/" }, "name": "Google News", "tags": [ "news", "google", "internal-hack-week" ], "verified_at": "2026-10-24" }, { "authors": [ "Torkel Ödegaard" ], "body": "## What it does\n\nGrafana renders time-series data from multiple backends into dashboards. It connects to Prometheus, InfluxDB, Elasticsearch, and dozens of other data sources through a plugin model.\n\n## The key insight\n\nTorkel Odegaard built it during a Christmas hackathon in 2013 because he wanted to see his own metrics. The architectural decision that made it succeed: separate the visualization layer from the data source. Any backend can be a plugin.\n\n## How it works\n\nGrafana has three main pieces: the panel, the data source, and the dashboard.\n\nA **panel** is a single visualization (a graph, a table, a stat). It knows how to draw itself but not where the data comes from. It asks a data source for a query result, then renders whatever comes back.\n\nA **data source** is a plugin that knows how to talk to one backend. Prometheus has one, InfluxDB has one, and you can write your own. Each implements the same interface: take a query, return time-series data in a standard format.\n\nA **dashboard** is a collection of panels with shared time ranges and variables. Change the time picker and every panel re-queries its data source with the new range.\n\nThe plugin model is the leverage. Adding support for a new database does not require changing any visualization code. You write one adapter, and every panel type works with it.\n\n## Build something similar\n\nStart with a web app that renders a chart from one data source. Define a simple interface: `query(range)` returns a list of `(timestamp, value)` pairs.\n\nAdd a second data source behind the same interface. The moment you have two, you have the core of Grafana. The third and tenth take the same effort as the second.\n\nThe hard part is not the charting. It is designing the query interface wide enough that different backends can implement it without contorting their natural query language.\n\n## Sources\n\nOrigin story and architecture are described in the Grafana documentary and podcast interviews linked in the evidence field.\n", "deep_dive": true, "description": "Open-source analytics and interactive visualization platform for metrics, logs, and dashboards.", "domain": "developer-tools", "event": { "name": "Christmas hackathon", "year": 2013 }, "evidence": [ { "supports": "Creator Torkel Ödegaard: \"it started as a Christmas hackathon\"", "url": "https://kubernetespodcast.com/episode/122-grafana" }, { "supports": "First commit December 25, 2013 by Torkel Ödegaard", "url": "https://grafana.com/blog/the-story-of-grafana-documentary-from-one-developers-dream-to-20-million-users-worldwide" } ], "hot": 60, "id": "grafana", "key_idea": "Separate the visualization layer from the data source so any backend can plug in.", "learn_from": "Solve your own pain point first, then abstract the interface when you add the second integration.", "links": { "demo": "https://play.grafana.org", "official": "https://grafana.com", "source": "https://github.com/grafana/grafana" }, "name": "Grafana", "tags": [ "observability", "dashboards", "open-source", "monitoring", "playable" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nGreetings from Earth is an interactive visualization that treats the ocean as a climate garden. Users explore the sea as a system that needs tending.\n\n## What makes it interesting\n\nIt won the Galactic Impact award at NASA Space Apps 2023. The metaphor does the teaching: a garden implies care and seasons, which a temperature chart does not.\n\n## What you can learn\n\nMetaphor is a compression scheme. A good one carries relationships the audience already understands.\n\n## Limitations of this analysis\n\nThe record does not link the visualization or describe the datasets behind it.\n", "description": "Interactive ocean-as-climate-garden visualization.", "domain": "climate", "event": { "award": "Galactic Impact", "name": "NASA Space Apps Challenge", "year": 2023 }, "evidence": [ { "supports": "Space Apps 2023 Galactic Impact award", "url": "https://www.spaceappschallenge.org/blog/the-nasa-space-apps-challenge-live-global-winners-announcement/" } ], "hot": 23, "id": "greetings-from-earth", "name": "Greetings from Earth", "tags": [ "climate", "visualization", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "authors": [ "Jared Hecht", "Steve Martocci" ], "body": "## What it is\n\nGroupMe is group messaging for phones. You create a group, add people, and everyone sees the same thread. The app runs on smartphones, and members without one can join over SMS.\n\n## The problem\n\nGroup texting in 2010 was a patchwork. Carrier group MMS behaved differently across devices, and adding someone mid-thread often broke the thread. A group that mixed smartphones and basic phones had no shared space.\n\n## How the experience works\n\nA message sent from the app reaches every member as a message in the group, and replies return to the same thread. The SMS path matters because it removes the requirement that everyone installs the app before the conversation can start.\n\n## What makes it interesting\n\nJared Hecht and Steve Martocci built it at the TechCrunch Disrupt New York hackathon in 2010. The product decision that carried it was treating SMS as a first-class client rather than a fallback.\n\n## What you can learn\n\nA communication product is only as useful as its least-equipped member. Supporting the oldest channel in the group can be the feature that makes the rest work.\n\n## Limitations of this analysis\n\nThe record does not document the gateway architecture, carrier handling, or delivery guarantees.\n", "description": "Free group messaging app that works over SMS and smartphone apps, built as a hackathon project.", "domain": "web-mobile", "event": { "name": "TechCrunch Disrupt New York hackathon", "organizer": "TechCrunch", "year": 2010 }, "evidence": [ { "supports": "Conceived at the TechCrunch Disrupt: New York hackathon", "url": "https://techcrunch.com/2010/08/25/groupme-born-at-techcrunch-disrupt-secures-funding-and-launches/" }, { "supports": "Created at a hackathon in May 2010; launched beta at TechCrunch Disrupt", "url": "https://www.businessinsider.com/what-is-groupme" }, { "supports": "Named as a company born from a TechCrunch Disrupt 2010 hackathon", "url": "https://en.wikipedia.org/wiki/Hackathon" } ], "hot": 10, "id": "groupme", "links": { "official": "https://groupme.com" }, "name": "GroupMe", "tags": [ "messaging", "sms", "mobile" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nG.R.O.W. is a mapping tool for wildfire recovery. It gives residents and planners a view of damage and recovery work in their area.\n\n## The problem\n\nAfter a fire, information is scattered across agencies and spreadsheets. People who lost property need one place to see what happened and what is next.\n\n## What makes it interesting\n\nIt won the Local Impact award at NASA Space Apps 2024. The entry targets a specific community rather than a global audience, which is why a local award fits.\n\n## What you can learn\n\nA recovery tool has to be usable by people under stress. Fewer layers and clearer labels beat more data.\n\n## Limitations of this analysis\n\nThe record does not link the tool or describe its data sources or deployment.\n", "description": "User-friendly GIS mapping tool for wildfire recovery and observation.", "domain": "civic-tech", "event": { "award": "Local Impact", "name": "NASA Space Apps Challenge", "year": 2024 }, "evidence": [ { "supports": "Space Apps 2024 Local Impact award", "url": "https://news.sciencex.in/nasa-international-space-apps-challenge-announces-2024-global-winners" } ], "hot": 25, "id": "grow-wildfire-recovery", "name": "G.R.O.W.", "tags": [ "wildfire", "gis", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nGROWBOT is a robot-assisted system for optimizing tree growth. It collects data about a tree and automates some of the care.\n\n## The problem\n\nTree planting gets attention; keeping saplings alive is the harder part. Growth depends on water, soil, and timing that a volunteer cannot monitor continuously.\n\n## What makes it interesting\n\nIt was named Hackathon Champion at robo.innovate 2025 in Munich, with a €2,500 prize. The project pairs a physical robot with a data loop rather than treating planting as a one-time act.\n\n## What you can learn\n\nA hardware project needs a maintenance story. A robot that tends trees has to survive the weather it is built for.\n\n## Limitations of this analysis\n\nThe record does not describe the robot, the sensors, or any field trial.\n", "description": "Robot-assisted tree growth optimization system; named Hackathon Champion at robo.innovate 2025 in Munich with a €2,500 prize.", "domain": "robotics", "event": { "award": "Hackathon Champion", "name": "robo.innovate Hackathon", "organizer": "MIRMI / Technical University of Munich", "year": 2025 }, "evidence": [ { "supports": "GROWBOT named Hackathon Champion at robo.innovate 2025 with €2,500 prize for robotic-assisted tree growth optimization via data collection and automated intervention", "url": "https://www.mirmi.tum.de/en/mirmi/news/article/roboinnovate-hackathon-2025-advancing-robotics-and-ai-solutions" } ], "hot": 35, "id": "growbot", "links": { "official": "https://www.mirmi.tum.de/en/mirmi/news/article/roboinnovate-hackathon-2025-advancing-robotics-and-ai-solutions" }, "name": "GROWBOT", "tags": [ "robo-innovate", "agriculture", "robotics", "hackathon-champion" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nHerCode Space is a NASA Space Apps 2025 winner in the Women and Space theme category.\n\n## What makes it interesting\n\nThe award recognizes a project built around women in space, a theme the competition added to widen who sees themselves in the data.\n\n## What you can learn\n\nA themed category is a curation decision. It tells entrants which problems the organizers think are under-addressed.\n\n## Limitations of this analysis\n\nThe record names the award but does not describe the project, its data, or its team. There is nothing here to evaluate beyond the recognition.\n", "description": "Space Apps 2025 winner in the Women/Space theme category.", "domain": "social", "event": { "name": "NASA Space Apps Challenge", "year": 2025 }, "evidence": [ { "supports": "Space Apps 2025 winner", "url": "https://www.spaceappschallenge.org/2025/awards/global-winners/" } ], "hot": 5, "id": "hercode-space", "name": "HerCode Space", "tags": [ "space", "diversity", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "authors": [ "Team Innovortex, LPU" ], "body": "## What it is\n\nHEXDOC is a robot that avoids collisions while moving through tunnels and caves. It targets construction and excavation work where the machine has to navigate without a clear line of sight.\n\n## The problem\n\nTunnelling equipment operates in confined, dusty spaces where a collision can injure a worker or damage the machine. A human spotter is not always possible.\n\n## What makes it interesting\n\nTeam Innovortex from LPU won the Hardware Edition of Smart India Hackathon 2023 in the Robotics and Drones category. The event asks teams to solve a stated industry problem, which is why the prototype is aimed at a specific job.\n\n## What you can learn\n\nA constrained environment is a good first market for autonomy. Tunnels remove the variables that make outdoor robots hard.\n\n## Limitations of this analysis\n\nThe record does not describe the sensors, the avoidance logic, or any field test.\n", "description": "Collision-avoidance robot for construction and excavation in tunnels and caves; won Smart India Hackathon 2023 Hardware Edition Student Innovation in Robotics and Drones.", "domain": "robotics", "event": { "award": "Hardware Edition Student Innovation winner (Robotics and Drones)", "name": "Smart India Hackathon", "organizer": "Government of India / AICTE", "year": 2023 }, "evidence": [ { "supports": "Team Innovortex from LPU won Smart India Hackathon 2023 Hardware Edition Student Innovation (Robotics and Drones) with HEXDOC, a collision-avoidance robot for tunnels and caves; 1,305 teams competed", "url": "https://happenings.lpu.in/lpu-innovortex-triumphs-smart-india-hackathon-2023" } ], "hot": 33, "id": "hexdoc", "links": { "official": "https://happenings.lpu.in/lpu-innovortex-triumphs-smart-india-hackathon-2023" }, "name": "HEXDOC", "tags": [ "smart-india-hackathon", "robot", "construction", "hardware" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nHipHop for PHP converts PHP source into C++ and compiles the result. Facebook ran its PHP web tier through it and reported roughly half the CPU usage for the same work.\n\n## The key insight\n\nFacebook did not want to rewrite the site. The team wanted the existing PHP codebase to run faster without a language migration. HipHop keeps PHP as the source language and moves the cost from request time to build time.\n\nThat trade is the whole idea. Compilation is slow and awkward, but it happens once. Interpretation happens on every request.\n\n## How it works\n\nThe compiler parses PHP, builds an abstract syntax tree, and generates C++ that calls a runtime library for dynamic behavior. PHP is dynamically typed, so the generated code cannot assume types. The runtime provides the operations that need them: array access, string conversion, method dispatch.\n\nStatic analysis recovers what it can. Where a variable's type is provable, the generated C++ uses a concrete type. Where it is not, the code falls back to a generic path. The ratio between the two determines how much speed the build actually buys.\n\nThe output is compiled with a normal C++ toolchain and linked against the HipHop runtime. Deployment ships a binary instead of a directory of scripts, which changes the release process as much as it changes performance.\n\nCompatibility is the long tail. PHP's behavior in edge cases is defined by its interpreter, and a reimplementation has to match it closely enough that production code does not change meaning. Each mismatch is a bug that only appears on a specific code path.\n\nThe project later gave way to HHVM, a virtual machine with a JIT. That design keeps the dynamic language and recovers speed at runtime, which removed the build step that made HipHop expensive to operate.\n\n## Build something similar\n\nTake a small language or a configuration format and compile it to a language you already run. Start with arithmetic, variables, and function calls.\n\nGenerate code, compile it, and compare the output against an interpreter for the same program. Then add one dynamic feature, such as a dictionary with mixed value types, and watch the generated code lose its static types.\n\nThat moment shows where the real work lives. The compiler is straightforward. Matching the dynamic semantics is the project.\n\n## Sources\n\nOrigin, author, and CPU savings are linked in the evidence field.\n", "deep_dive": true, "description": "PHP to C++ compiler that reduced CPU usage by approximately 50%. Open-sourced in February 2010.", "domain": "developer-tools", "event": { "name": "Facebook Hackathon", "organizer": "Facebook", "year": 2010 }, "evidence": [ { "supports": "Facebook hackathon origin, Haiping Zhao, CPU savings", "url": "https://techcrunch.com/2010/02/02/facebook-hiphop-presentation/" } ], "hot": 30, "id": "hiphop-for-php", "key_idea": "Translate PHP into C++ ahead of time, so the same code runs without paying interpreter cost per request.", "learn_from": "A code generator plus a compatibility layer can buy years of headroom while you build the real runtime.", "name": "HipHop for PHP", "notes": "Built by Haiping Zhao", "tags": [ "php", "compiler", "facebook", "internal-hack-week" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nHollow Knight is a metroidvania set in a ruined underground kingdom. You play a small knight with a nail, explore interconnected rooms, and unlock movement abilities that open routes you could see but not reach earlier.\n\n## The key insight\n\nThe project started as a Ludum Dare 27 entry in 2013, built in a weekend around the theme \"ten seconds.\" What survived from that build was not its content. It was the feel: a small figure moving through a dark, hand-drawn space.\n\nTeam Cherry kept that and rebuilt the rest over several years. The jam version proved the tone cheaply. The commercial game needed a much larger map, a combat system, and a save structure, and none of those fit in a weekend.\n\n## How it works\n\nThe map is a graph. Each room is a node, each exit is an edge, and most edges are gated by either an ability or a key item. Designers place gates so that the player passes a locked route early and remembers it. Later, a new ability turns the remembered route into progress.\n\nCombat is built on spacing and timing rather than combos. The nail has a short range, the dash moves a fixed distance, and healing takes time during which the player is vulnerable. Every fight is a negotiation between dealing damage and buying the moment to heal.\n\nThe charm system is a loadout constraint. Charms modify health, damage, movement, or summon behavior, but you can only equip as many as your notch count allows. The constraint turns a large set of upgrades into a build decision instead of a checklist.\n\nSaving happens at benches. A bench restores health and respawns enemies, so progress and risk are tied to a physical location rather than a menu.\n\n## Build something similar\n\nStart with one room and one movement verb. Make the character move, jump, and attack. Tune the acceleration, the jump arc, and the attack recovery until moving through an empty room feels good on its own.\n\nThen build three rooms with one gate. Put an item behind the gate that the player can see but not reach. Add the ability that opens it. If the return trip feels like a reward rather than a chore, the core loop works.\n\nAdd a bench before adding a boss. Saving is the pacing device that makes the rest tolerable.\n\n## Sources\n\nJam origin and development history are linked in the evidence field.\n", "deep_dive": true, "description": "Metroidvania set in a ruined insect kingdom. Team Cherry's jam prototype seeded the full game.", "domain": "games", "event": { "award": "Theme \"Ten seconds\" entry", "name": "Ludum Dare 27", "year": 2013 }, "evidence": [ { "supports": "Ludum Dare 27 origin and Team Cherry development history", "url": "https://gamemaker.io/en/blog/best-game-jam-games" } ], "hot": 65, "id": "hollow-knight", "key_idea": "Keep the tone and movement from the jam build, and rebuild everything around them at full scale.", "learn_from": "A jam is a test of feel, not a test of content. Ship the smallest thing that proves the mood, then start over on systems.", "links": { "official": "https://hollowknightgame.com/", "source": "https://github.com/teamcherry/HollowKnight" }, "name": "Hollow Knight", "tags": [ "game-jam", "metroidvania", "indie", "unity", "open-source" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nHoneyPause is a marketplace for security reviews of on-chain projects. It connects protocol teams with auditors and white-hat researchers.\n\n## The problem\n\nSmart contracts hold funds from the day they deploy. A team that wants a review has to find a reviewer, agree on scope, and trust the result, often under time pressure.\n\n## What makes it interesting\n\nIt won Best of ETHDenver 2024. The project is infrastructure for trust rather than another application, which is a harder pitch in a hackathon full of demos.\n\n## What you can learn\n\nA two-sided marketplace has to solve the cold start for the scarce side. Here that is the auditor's time, not the protocol's demand.\n\n## Limitations of this analysis\n\nThe record does not describe the escrow model, the reputation system, or whether any audits were completed.\n", "description": "White-hat security marketplace for on-chain projects, connecting auditors with protocol teams.", "domain": "fintech", "event": { "award": "Best of ETHDenver", "name": "ETHDenver", "year": 2024 }, "evidence": [ { "supports": "ETHDenver 2024 Best of ETHDenver award", "url": "https://unchainedcrypto.com/ethdenvers-2024-hackathon-onchain-security-ai-and-gaming-fun-dominate-the-winning-projects/" } ], "hot": 25, "id": "honeypause", "name": "HoneyPause", "tags": [ "web3", "security", "ethereum" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nImmersive Rehab is a virtual reality platform for rehabilitation. It targets neurological conditions where movement practice is part of recovery.\n\n## The problem\n\nRepetitive exercises are boring, and bored patients stop. VR can make the repetition tolerable and give the therapist data.\n\n## What makes it interesting\n\nIt was featured by Red Bull Basement. The project sits at the intersection of two hard markets, clinical validation and consumer hardware, which is why so few rehabilitation products reach patients.\n\n## What you can learn\n\nIn clinical VR, the therapist is a user too. A product that ignores their workflow will not be adopted, however good the headset demo is.\n\n## Limitations of this analysis\n\nThe record does not describe the exercises, the clinical evidence, or the regulatory path.\n", "description": "Virtual reality rehabilitation platform for neurological conditions.", "domain": "healthcare", "event": { "name": "Red Bull Basement" }, "evidence": [ { "supports": "Red Bull Basement featured project", "url": "https://www.redbull.com/my-en/videos/soj-gamayon-in-la-and-silicon-valley" } ], "hot": 0, "id": "immersive-rehab", "name": "Immersive Rehab", "tags": [ "vr", "rehabilitation", "red-bull-basement" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nImmortal's Way is a game first created at Global Game Jam 2024 and later expanded into a full Steam release.\n\n## What makes it interesting\n\nThe jam prototype established the setting and the core loop. The commercial version kept the concept and built the content around it.\n\n## What you can learn\n\nA jam is a cheap way to test whether a world is worth spending a year in. If the team still likes it on Monday, the idea has legs.\n\n## Limitations of this analysis\n\nThe record does not describe the mechanics or how much of the jam build survived.\n", "description": "Game first created at Global Game Jam 2024 that was later expanded into a full commercial release on Steam.", "domain": "games", "event": { "name": "Global Game Jam 2024", "organizer": "Global Game Jam", "year": 2024 }, "evidence": [ { "supports": "GGJ lists Immortal's Way as originally created at Global Game Jam 2024, later evolved into a full Steam game", "url": "https://globalgamejam.org/ggj-success-stories" } ], "hot": 15, "id": "immortals-way", "links": { "official": "https://store.steampowered.com/app/2986350/Immortals_Way/" }, "name": "Immortal's Way", "tags": [ "global-game-jam", "commercial", "indie" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nInstaDApp is a middleware layer for decentralized finance. It gives users one interface for lending protocols such as MakerDAO instead of separate positions in each.\n\n## The problem\n\nEach DeFi protocol has its own interface and its own way of managing collateral. A user with positions in several places has to watch all of them separately.\n\n## What makes it interesting\n\nBrothers Sowmay and Samyak Jain built it at ETHIndia 2018 in Bengaluru and won the hackathon. The product became a way to manage a portfolio of positions rather than a single protocol.\n\n## What you can learn\n\nAbstraction over competing protocols is a durable position. The user wants the outcome, not loyalty to one venue.\n\n## Limitations of this analysis\n\nThe record does not describe the smart contract architecture or how the platform handled risk across protocols.\n", "description": "Decentralized finance middleware platform that simplifies interactions with lending protocols like MakerDAO through a unified interface.", "domain": "fintech", "event": { "name": "ETHIndia", "year": 2018 }, "evidence": [ { "supports": "Quote from Sowmay Jain, Founder of InstaDApp: 'The fact that our hack [InstaDApp] saw the light of the day is a great signal about the success of the hackathon'", "url": "https://online.ethindia.co/" }, { "supports": "InstaDApp started as a hack at ETHIndia in Bengaluru in 2018 by brothers Sowmay and Samyak Jain", "url": "https://www.financialexpress.com/market/instadapp-how-two-indian-brothers-in-their-20s-put-a-blockchain-startup-on-the-world-map-2261761" }, { "supports": "They won the ETHIndia hackathon where they first created the Instadapp platform", "url": "https://www.globalindian.com/youth/samyak-and-sowmay-jain-disrupting-decentralised-finance-through-instadapp" } ], "hot": 1, "id": "instadapp", "links": { "official": "https://instadapp.io" }, "name": "InstaDApp", "tags": [ "defi", "blockchain", "ethereum" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nInstructAI is a generative-AI tutor for programming. It helps a student work through code rather than handing over the answer.\n\n## The problem\n\nIntroductory programming is where students get stuck, and one teaching assistant cannot watch every editor. A tutor that responds in the moment changes how long a beginner stays blocked.\n\n## What makes it interesting\n\nGeorgia Tech students built it and won the 2024 Klaus Startup Challenge, which came with a $100,000 seed offer at a $1 million valuation.\n\n## What you can learn\n\nA university is a controlled market for an education product. You can test with real courses before you sell to a district.\n\n## Limitations of this analysis\n\nThe record does not describe the tutoring model, its accuracy, or any learning-outcome study.\n", "description": "Generative-AI programming tutor built by Georgia Tech students; won the 2024 Klaus Startup Challenge with a $100K seed offer at a $1M valuation.", "domain": "education", "event": { "award": "winner", "name": "Klaus Startup Challenge", "organizer": "Georgia Tech College of Computing", "year": 2024 }, "evidence": [ { "supports": "InstructAI won the 2024 Klaus Startup Challenge at Georgia Tech College of Computing; offered $100,000 seed funding for 10% equity ($1M valuation)", "url": "https://sci.cc.gatech.edu/external-news/students-win-klaus-startup-challenge-dive-5b-market-ai-driven-teaching-tool" } ], "hot": 25, "id": "instructai", "links": { "official": "https://www.cc.gatech.edu/news/students-win-klaus-startup-challenge-dive-5b-market-ai-driven-teaching-tool" }, "name": "InstructAI", "tags": [ "georgia-tech", "ai", "edtech", "student-hackathon" ], "verified_at": "2026-10-24" }, { "authors": [ "Arnav Chintawar", "Dhruv Roongta", "Sahibpreet Singh" ], "body": "## What it is\n\nJARVIS is an assistant app for blind and visually impaired users. It uses computer vision to describe surroundings and non-verbal cues.\n\n## The problem\n\nA white cane tells you what is directly in front of you. It does not tell you whether someone is smiling, whether a room is crowded, or who is speaking.\n\n## What makes it interesting\n\nGeorgia Tech first-year computer science students built it at Cal Hacks 10.0 in 2023 and won Best Hack for Health, along with prizes for Zilliz and GitHub. The project pairs a vision model with a conversational interface.\n\n## What you can learn\n\nAccessibility tools benefit from a familiar output channel. Speech is already how a phone assistant works, so the new capability fits an existing habit.\n\n## Limitations of this analysis\n\nThe record does not report recognition accuracy, latency, or testing with blind users.\n", "description": "AI assistant app that uses computer vision to describe surroundings and non-verbal cues for blind and visually impaired users; won Best Hack for Health at Cal Hacks 10.0.", "domain": "accessibility", "event": { "award": "Best Hack for Health", "name": "Cal Hacks 10.0", "organizer": "UC Berkeley HKN", "year": 2023 }, "evidence": [ { "supports": "Georgia Tech first-year CS majors built JARVIS at Cal Hacks 10.0 and won Best Hack for Health plus Best Use of Zilliz and Best Use of GitHub", "url": "https://news.gatech.edu/news/2023/12/11/cs-majors-win-best-hack-health-app-visually-impaired" }, { "supports": "College of Computing writeup: JARVIS uses open-source AI, ML and computer vision to give visually impaired users situational awareness; inspired by a visually impaired family member", "url": "https://www.cc.gatech.edu/news/cs-majors-win-best-hack-health-app-visually-impaired" } ], "hot": 23, "id": "jarvis", "links": { "official": "https://news.gatech.edu/news/2023/12/11/cs-majors-win-best-hack-health-app-visually-impaired" }, "name": "JARVIS", "tags": [ "ai", "computer-vision", "accessibility", "cal-hacks" ], "verified_at": "2026-10-24" }, { "authors": [ "Brinlee Kidd", "Sylvia Lopez" ], "body": "## What it is\n\nJotted is an automated study tool. It turns material into notes and review aids for students.\n\n## What makes it interesting\n\nArizona State University students Brinlee Kidd and Sylvia Lopez won the Red Bull Basement 2021 Global Final in Istanbul with it. The competition rewards a clear student problem and a demo that a non-technical judge can understand.\n\n## What you can learn\n\nA student-built tool has an advantage: the builders are the users. That makes the first version honest about what actually helps.\n\n## Limitations of this analysis\n\nThe record does not describe the product's features, its accuracy, or what happened after the competition.\n", "description": "Automated study tool by Arizona State University students Brinlee Kidd and Sylvia Lopez; won the Red Bull Basement 2021 Global Final in Istanbul.", "domain": "education", "event": { "award": "Global Final winner", "name": "Red Bull Basement", "organizer": "Red Bull", "year": 2021 }, "evidence": [ { "supports": "Arizona State University students Brinlee Kidd and Sylvia Lopez won Red Bull Basement 2021 Global Final in Istanbul with Jotted, an automated study tool; 4,000+ applications that year", "url": "https://www.prnewswire.com/news-releases/automated-study-tool-wins-best-idea-at-innovation-program-for-university-students-301513791.html" } ], "hot": 23, "id": "jotted", "links": { "official": "https://www.prnewswire.com/news-releases/automated-study-tool-wins-best-idea-at-innovation-program-for-university-students-301513791.html" }, "name": "Jotted", "tags": [ "red-bull-basement", "edtech", "student-innovation", "study-tools" ], "verified_at": "2026-10-24" }, { "authors": [ "Steel Crate Games" ], "body": "## What it is\n\nKeep Talking and Nobody Explodes is a cooperative bomb-defusal game. One player sees a bomb in VR; the others hold a manual and cannot see it.\n\n## The problem\n\nCo-op games usually give every player the same information, which makes coordination a matter of reflexes. This one makes coordination the entire game by splitting the information.\n\n## How the experience works\n\nThe defuser describes what is on the bomb: a serial number, a symbol, a colored wire. The experts find the matching rule in the manual and read back instructions. Neither side can finish alone.\n\n## What makes it interesting\n\nIt began as a Global Game Jam 2015 prototype and became one of the best-selling VR titles on Steam. The manual is printable paper, so the game supports any number of players without extra hardware.\n\n## What you can learn\n\nAsymmetric information is a cheap source of depth. Before adding mechanics, check whether splitting what players know creates the tension you want.\n\n## Limitations of this analysis\n\nThe record does not describe the bomb-generation logic or how the VR and non-VR versions differ.\n", "description": "Cooperative VR bomb-defusal game born from a Global Game Jam 2015 prototype; became one of the best-selling VR titles on Steam.", "domain": "games", "event": { "award": "jam prototype", "name": "Global Game Jam", "organizer": "Global Game Jam", "year": 2015 }, "evidence": [ { "supports": "Keep Talking and Nobody Explodes Steam page; originated from a Global Game Jam 2015 prototype and became a defining VR co-op title", "url": "https://store.steampowered.com/app/341800/Keep_Talking_and_Nobody_Explodes/" } ], "hot": 31, "id": "keep-talking", "links": { "official": "https://store.steampowered.com/app/341800/Keep_Talking_and_Nobody_Explodes/" }, "name": "Keep Talking and Nobody Explodes", "tags": [ "game-jam", "vr", "co-op", "steam", "ggj" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nLandsat Connect is a web app for following Landsat satellites. It shows surface reflectance data and notifies users when a satellite passes over an area.\n\n## The problem\n\nLandsat data is open but not approachable. A researcher or land manager has to know which scene to request and when it was captured.\n\n## What makes it interesting\n\nIt won Best Mission Concept at NASA Space Apps 2024. The pass notification turns an archive into a subscription, which is the difference between data being available and being used.\n\n## What you can learn\n\nNotifications are a product feature, not a nicety. A dataset that changes on a schedule is more useful when it tells you the schedule.\n\n## Limitations of this analysis\n\nThe record does not link the app or describe the data pipeline.\n", "description": "Web app tracking Landsat satellites with surface reflectance data and pass notifications.", "domain": "science", "event": { "award": "Best Mission Concept", "name": "NASA Space Apps Challenge", "year": 2024 }, "evidence": [ { "supports": "Space Apps 2024 Best Mission Concept award", "url": "https://news.sciencex.in/nasa-international-space-apps-challenge-announces-2024-global-winners" } ], "hot": 25, "id": "landsat-connect", "name": "Landsat Connect", "tags": [ "earth-observation", "web-mobile", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nLaunchRock built pre-launch landing pages that collected email addresses before a product shipped.\n\n## The problem\n\nA team with no product still needs a way to measure interest. A landing page with a signup form is the cheapest test of demand.\n\n## What makes it interesting\n\nIt placed second at Startup Weekend Philadelphia in 2011, raised seed funding, and was acquired about three years after launch. The product was a page and a form, which is a small surface for a company.\n\n## What you can learn\n\nSell the waitlist before you build the product. If nobody signs up, you saved a year.\n\n## Limitations of this analysis\n\nThe record does not describe the page builder, the pricing, or the acquisition.\n", "description": "Pre-launch landing page builder for collecting emails before product launch. Seed funded, acquired approximately three years after launch.", "domain": "web-mobile", "event": { "award": "2nd place", "name": "Startup Weekend Philadelphia", "year": 2011 }, "evidence": [ { "supports": "Startup Weekend Philadelphia 2011 second place and subsequent acquisition", "url": "https://www.inc.com/articles/201111/companies-that-launched-in-three-days-startup-weekend.html" } ], "hot": 30, "id": "launchrock", "name": "LaunchRock", "tags": [ "marketing", "landing-page", "startup-weekend" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nLAVA AQUA X is a portable electric washing machine designed to use less water. It was built by the team LYLO.\n\n## The problem\n\nLaundry is a large share of household water use, and portable machines often trade efficiency for size.\n\n## What makes it interesting\n\nIt won the Global Workshop at Red Bull Basement 2020. The project treats water use as a design constraint from the start rather than an afterthought.\n\n## What you can learn\n\nA sustainability claim needs a measurable baseline. The winning pitch is stronger when it states liters per load, not \"less water\".\n\n## Limitations of this analysis\n\nThe record does not include the water savings, the prototype status, or any testing.\n", "description": "Portable electric washing machine designed for water conservation. Built by team LYLO.", "domain": "hardware", "event": { "award": "Global Workshop winner", "name": "Red Bull Basement", "year": 2020 }, "evidence": [ { "supports": "Red Bull Basement 2020 Global Workshop winner", "url": "https://www.redbull.com/int-en/event-series/basement/red-bull-basement-past-winners" } ], "hot": 23, "id": "lava-aqua-x", "name": "LAVA AQUA X", "tags": [ "water", "sustainability", "red-bull-basement" ], "verified_at": "2026-10-24" }, { "authors": [ "Antoine Balaresque", "Henry Bradlow" ], "body": "## What it is\n\nLily Robotics proposed a throw-and-shoot camera drone that follows its owner. The first prototype was built in the UC Berkeley Robotics Laboratory.\n\n## The problem\n\nA follow-me camera lets one person film themselves doing something active. Handheld footage and tripods both fail when the subject is moving.\n\n## What makes it interesting\n\nThe company took about $34 million in pre-orders, then shut down in January 2017 after failing to raise manufacturing funds. The gap between a lab demo and a shippable product is the story.\n\n## What you can learn\n\nA pre-order is a promise with a date attached. Selling before the manufacturing path is proven converts customer enthusiasm into a liability.\n\n## Limitations of this analysis\n\nThe record does not cover the fraud allegations in detail, the refund process, or the engineering state at shutdown.\n", "description": "UC Berkeley robotics-lab prototype for a throw-and-shoot follow-me camera drone that drew huge pre-orders, then collapsed amid fraud allegations and bankruptcy.", "domain": "robotics", "evidence": [ { "supports": "Founded 2013 by UC Berkeley alums Antoine Balaresque and Henry Bradlow; first prototype built in the UC Berkeley Robotics Laboratory using a Raspberry Pi and Arduino", "url": "https://en.wikipedia.org/wiki/Lily_Robotics" }, { "supports": "Lily Robotics shut down in January 2017 despite about $34 million in pre-orders after failing to raise manufacturing funds", "url": "https://fortune.com/2017/01/12/drone-startup-abruptly-shuts" }, { "supports": "Long-form account of Lily's rise and fall from a Berkeley lab prototype to a crashed hype cycle", "url": "https://www.wired.com/story/the-drone-company-that-fell-to-earth" }, { "supports": "Bankruptcy filing details demand shock and refund plans after tens of thousands of prepaid orders", "url": "https://www.siliconvalley.com/2017/02/28/shuttered-drone-startup-lily-robotics-details-failure-refund-plans-bankruptcy-filing/amp" } ], "hot": 0, "id": "lily-robotics", "links": { "official": "https://en.wikipedia.org/wiki/Lily_Robotics" }, "name": "Lily Robotics", "tags": [ "drones", "hardware", "cautionary-tale", "berkeley" ], "verified_at": "2026-10-24" }, { "authors": [ "Four Quarters" ], "body": "## What it is\n\nLoop Hero is an idle role-playing game. A hero walks an endless loop while the player places terrain and enemies around the path.\n\n## How the experience works\n\nThe hero moves automatically and fights what the loop produces. The player's decisions happen outside the walk: which tiles to place, which resources to keep, when to return to camp.\n\n## What makes it interesting\n\nThe first version was a Ludum Dare 45 entry in 2019 that only had the walking loop. Four Quarters built the rest and released through Devolver.\n\n## What you can learn\n\nA prototype can be one system. If watching a character walk is already interesting, the game has a foundation.\n\n## Limitations of this analysis\n\nThe record does not describe the tile system or how the team balanced the idle loop.\n", "description": "Idle-RPG about a hero walking an endless loop against monsters, grown from a non-playable Ludum Dare 2019 prototype into Four Quarters' Devolver-published hit.", "domain": "games", "event": { "name": "Ludum Dare 45", "organizer": "Ludum Dare", "year": 2019 }, "evidence": [ { "supports": "First version of Loop Hero was a 2019 Ludum Dare entry that only had the walking loop; the unfinished jam build sparked the full game", "url": "https://www.vice.com/en/article/how-the-loop-hero-devs-pitched-a-hit-game-thats-impossible-to-describe/" }, { "supports": "Original Ludum Dare 45 submission page for Loop a Hero / Loop Hero prototype", "url": "https://ldjam.com/events/ludum-dare/45/loopathero" } ], "hot": 26, "id": "loop-hero", "links": { "demo": "https://ldjam.com/events/ludum-dare/45/loopathero", "official": "https://fourquarterslab.com" }, "name": "Loop Hero", "tags": [ "ludum-dare", "rpg", "indie", "prototype", "playable" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nLoot Crate was a subscription box for geek and gamer merchandise. Each month shipped a themed collection of items.\n\n## The problem\n\nFandom merchandise was scattered across stores and conventions. A curated box turns discovery into a recurring purchase.\n\n## What makes it interesting\n\nIt started at Startup Weekend Los Angeles in 2012 and grew to millions of subscribers. The business depended on licensing and logistics, not on software.\n\n## What you can learn\n\nA subscription is a promise about next month. The churn risk sits in the second box, not the first.\n\n## Limitations of this analysis\n\nThe record does not cover the box economics, the licensing deals, or the company's later decline.\n", "description": "Geek and gamer subscription box service with millions of subscribers.", "domain": "web-mobile", "event": { "name": "Startup Weekend Los Angeles", "year": 2012 }, "evidence": [ { "supports": "Startup Weekend LA 2012 origin", "url": "https://medium.com/startup-weekend-derby/5-startups-who-used-startup-weekend-as-a-launchpad-for-success-fa380463177e" } ], "hot": 10, "id": "loot-crate", "links": { "official": "https://www.lootcrate.com/" }, "name": "Loot Crate", "tags": [ "subscription", "retail", "startup-weekend" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nLovers in a Dangerous Spacetime is a couch co-op game. Players move around a circular ship and man its turrets, engines, and shields.\n\n## How the experience works\n\nThe ship has stations, and there are more stations than players. Coordination means deciding who leaves which post and when, which turns movement inside the ship into the puzzle.\n\n## What makes it interesting\n\nIt started at Global Game Jam 2012 in Toronto and later released on major consoles. The jam version established the shared-ship idea.\n\n## What you can learn\n\nA co-op design is interesting when players cannot do everything at once. Scarcity of attention is the mechanic.\n\n## Limitations of this analysis\n\nThe record does not describe the ship systems or the porting work.\n", "description": "Couch co-op space brawler where players man turrets on a circular ship. Released on all major consoles.", "domain": "games", "event": { "name": "Global Game Jam 2012", "organizer": "Toronto site" }, "evidence": [ { "supports": "GGJ 2012 Toronto origin and commercial release", "url": "https://globalgamejam.org/ggj-success-stories" } ], "hot": 10, "id": "lovers-in-a-dangerous-spacetime", "links": { "official": "https://loversinadangerousspacetime.com/" }, "name": "Lovers in a Dangerous Spacetime", "tags": [ "game-jam", "co-op", "indie" ], "verified_at": "2026-10-24" }, { "authors": [ "1 Simple Idea" ], "body": "## What it is\n\nLullaby of Life is a musical exploration game. The player uses sound and rhythm to move through its world.\n\n## What makes it interesting\n\nIt began as a Global Game Jam 2016 entry and later shipped on Apple Arcade and Steam. The jam scope suited a mechanic that is easier to feel than to explain.\n\n## What you can learn\n\nA mechanic built on sound needs a short demo. Players understand it in seconds or not at all.\n\n## Limitations of this analysis\n\nThe record does not describe the interaction or how the game expanded beyond the jam.\n", "description": "Musical exploration game that started as a Global Game Jam prototype and later shipped on Apple Arcade and Steam.", "domain": "games", "event": { "award": "jam prototype", "name": "Global Game Jam", "organizer": "Global Game Jam", "year": 2016 }, "evidence": [ { "supports": "Lullaby of Life began as a Global Game Jam 2016 entry before being expanded into an Apple Arcade and Steam title", "url": "https://www.gamedeveloper.com/design/how-a-global-game-jam-prototype-became-lullaby-of-life" } ], "hot": 31, "id": "lullaby-of-life", "links": { "official": "https://apps.apple.com/app/lullaby-of-life/id1503210927" }, "name": "Lullaby of Life", "tags": [ "game-jam", "music", "puzzle", "apple-arcade" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nLunarTech Ensemble is a website and immersive game that visualizes Apollo lunar seismic data.\n\n## The problem\n\nThe Apollo seismometers recorded moonquakes for years, but the data sits in archives that few people explore.\n\n## What makes it interesting\n\nIt won Best Use of Science at NASA Space Apps 2023. Turning a seismic record into something a person can move through is a way to make an old dataset legible.\n\n## What you can learn\n\nOld missions still hold unused data. A new interface can be the contribution, without a new instrument.\n\n## Limitations of this analysis\n\nThe record does not link the experience or describe how the seismic data is mapped.\n", "description": "Website and immersive game visualizing Apollo lunar seismic data.", "domain": "science", "event": { "award": "Best Use of Science", "name": "NASA Space Apps Challenge", "year": 2023 }, "evidence": [ { "supports": "Space Apps 2023 Best Use of Science award", "url": "https://www.spaceappschallenge.org/blog/the-nasa-space-apps-challenge-live-global-winners-announcement/" } ], "hot": 23, "id": "lunartech-ensemble", "name": "LunarTech Ensemble", "tags": [ "astronomy", "visualization", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nMedLink is a healthcare connectivity app built at San Francisco State University's civic hackathon in 2014. It was featured at a White House Champions of Change event for open government and civic hacking.\n\n## The problem\n\nHealth information is split between clinics, insurers, and patients. A civic hackathon is one way to prototype a bridge when the institutions cannot agree on one.\n\n## What makes it interesting\n\nThe White House recognition is unusual for a student hackathon project. It shows that a small prototype can serve as a policy story, not just a product.\n\n## What you can learn\n\nCivic technology often succeeds as a demonstration. The prototype makes an argument that procurement cannot.\n\n## Limitations of this analysis\n\nThe record does not describe the app's features, its data sources, or whether it was deployed.\n", "description": "Healthcare connectivity app developed at SF State's civic hackathon; featured at a White House Champions of Change event for open government and civic hacking.", "domain": "healthcare", "event": { "award": "White House featured project", "name": "San Francisco Day of Civic Hacking", "organizer": "Code for America / SF State", "year": 2014 }, "evidence": [ { "supports": "MedLink, developed at SF State's civic hackathon, was featured at a White House Champions of Change of Open Government and Civic Hacking event", "url": "https://commons.sfsu.edu/hackathon-sf-state" } ], "hot": 20, "id": "medlink", "links": { "official": "https://commons.sfsu.edu/hackathon-sf-state" }, "name": "MedLink", "tags": [ "civic-hackathon", "healthcare", "open-government" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nMetaloot is a launchpad that connects games to the Flow blockchain. It gives game teams a way to launch on-chain assets and reach players.\n\n## What makes it interesting\n\nIt was named a winner among ten finalists at ETHGlobal Bangkok 2024, competing for a $750,000 prize pool. The project targets game studios rather than end users, which is a different distribution problem.\n\n## What you can learn\n\nInfrastructure for creators only works if the creators already have players. The launchpad is a multiplier, not a source of demand.\n\n## Limitations of this analysis\n\nThe record does not describe the launch mechanics, the Flow integration, or any games that used it.\n", "description": "Launchpad connecting games to the Flow blockchain; named a top winner at the ETHGlobal Bangkok 2024 hackathon following Devcon.", "domain": "fintech", "event": { "award": "finalist winner", "name": "ETHGlobal Bangkok", "organizer": "ETHGlobal", "year": 2024 }, "evidence": [ { "supports": "Metaloot named a winner among 10 ETHGlobal Bangkok 2024 finalists; a launchpad built to connect games to Flow, competing for a $750,000 prize pool", "url": "https://www.theblock.co/amp/post/326975/ethglobal-hackathons-10-finalists-showcase-ai-agents-crypto-games-dao-tools-and-more" } ], "hot": 25, "id": "metaloot", "links": { "official": "https://ethglobal.com/showcase" }, "name": "Metaloot", "tags": [ "ethglobal", "blockchain", "gaming", "flow" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nImmersive Reader is an accessibility feature in Microsoft Office. It reads text aloud, breaks it into syllables, and adjusts spacing and contrast for readers with dyslexia and other learning differences.\n\n## The problem\n\nStandard document views assume a reader who can track lines and decode text easily. Small typographic changes can make the difference between reading and giving up.\n\n## What makes it interesting\n\nIt started as Learning Tools and won a Microsoft company-wide hackathon in 2015, beating more than 3,000 other entrants. It later shipped inside Office, where millions of people already read.\n\n## What you can learn\n\nDistribution decides whether an accessibility feature helps anyone. Building it into the tool people already use beats shipping a separate app.\n\n## Limitations of this analysis\n\nThe record does not describe the reading science behind the feature or its measured impact.\n", "description": "Accessibility tool built into Microsoft Office that helps people with dyslexia and learning differences read and write more effectively.", "domain": "accessibility", "event": { "name": "Microsoft company-wide hackathon", "year": 2015 }, "evidence": [ { "supports": "Winning hackathon project now helping millions; originated as Learning Tools within Microsoft", "url": "https://news.microsoft.com/features/winning-hackathon-project-is-now-helping-millions-of-people-worldwide-to-read-and-write" }, { "supports": "Immersive Reader team's project won the Hackathon, beating over 3,000 other entrants", "url": "https://www.microsoft.com/en-us/research/group/customer-insights-research/articles/immersive-reader-creates-more-inclusive-classrooms" } ], "hot": 11, "id": "microsoft-immersive-reader", "links": { "official": "https://www.microsoft.com/en-us/education/immersive-reader" }, "name": "Microsoft Immersive Reader", "tags": [ "accessibility", "education", "inclusion" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nMuseum Wakes is a narrative mystery experienced by voice inside a museum. It turns a collection into an interactive story.\n\n## How the experience works\n\nThe visitor speaks with the guide and moves through the galleries. The story draws on the museum's own collection data, so the exhibits are the set.\n\n## What makes it interesting\n\nIt was built at the Gemini 3 NYC Hackathon in February 2026 and uses 8,390 artifacts from the Met's Egyptian Wing through the Met Open API, with live generative responses.\n\n## What you can learn\n\nA museum already has the content and the space. The hack is the layer that makes a visitor look at both differently.\n\n## Limitations of this analysis\n\nThe record does not describe how the generated narration is kept factually grounded. That is the main risk in a museum context.\n", "description": "AI-powered narrative mystery transforming the Met's Egyptian Wing into an interactive voice experience; built at the Gemini 3 NYC Hackathon in February 2026.", "domain": "media-arts", "event": { "award": "gallery project", "name": "Gemini 3 NYC Hackathon", "organizer": "Cerebral Valley", "year": 2026 }, "evidence": [ { "supports": "Museum Wakes built at Gemini 3 NYC Hackathon Feb 28 2026; transforms Met Egyptian Wing using 8,390 artifacts via the Met Open API with live generative voice dialogue", "url": "https://cerebralvalley.ai/e/gemini-3-nyc-hackathon/hackathon/gallery/16" } ], "hot": 35, "id": "museum-wakes", "links": { "official": "https://museum-wakes.onrender.com/" }, "name": "Museum Wakes", "tags": [ "gemini", "museum", "voice-ai", "narrative", "hackathon" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nMushroom 11 is a platformer where the player controls a spreading fungal organism instead of a character.\n\n## The problem\n\nPlatformers are built on a body with a fixed shape. It jumps, it fits through gaps, it takes damage. A blob with no fixed shape needs different verbs.\n\n## How the experience works\n\nYou shape the organism to move it through spaces, growing one side and losing the other. Progress comes from managing the mass rather than from timing a jump.\n\n## What makes it interesting\n\nIt started at Global Game Jam 2012 at NYU and was shortlisted for an IGF award for design. The jam constrained the team to a mechanic that could be understood without a character or a story.\n\n## What you can learn\n\nRemove the body and the genre's assumptions go with it. A single unusual verb can carry a whole game if the level design keeps asking new questions of it.\n\n## Limitations of this analysis\n\nThe sources establish the origin and the award shortlist. The growth and destruction rules, and how the team tuned them, are not documented in the linked material.\n", "description": "Fungal transformation platformer where the player is a growing, splitting organism. IGF Excellence in Design shortlist.", "domain": "games", "event": { "name": "Global Game Jam 2012", "organizer": "NYU site" }, "evidence": [ { "supports": "GGJ 2012 NYU origin and IGF recognition", "url": "https://globalgamejam.org/ggj-success-stories" } ], "hot": 10, "id": "mushroom-11", "links": { "official": "https://mushroom11game.com/" }, "name": "Mushroom 11", "tags": [ "game-jam", "platformer", "indie" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nThis is a web-app prototype that connects smallholder farmers with local buyers and monitors agricultural land use with satellite imagery. It was built at NASA Space Apps Colombia in 2021.\n\n## The problem\n\nA small farm sells into a local chain with little visibility. Satellite land-use data can show what is growing and where, which helps both the farmer and the buyer plan.\n\n## What makes it interesting\n\nThe prototype was adopted by Colombia's Ministry of Agriculture. A weekend build reaching a ministry is rare, and it happened because the team solved a problem the ministry already had.\n\n## What you can learn\n\nPublic adoption is a distribution channel. If a government body can use the prototype, it can also fund the next version.\n\n## Limitations of this analysis\n\nThe record does not describe the adoption terms, the number of farmers reached, or the current state of the tool.\n", "description": "Web-app prototype connecting smallholder farmers with local buyers and monitoring agricultural land-use via satellite imagery.", "domain": "climate", "event": { "name": "NASA Space Apps Challenge", "year": 2021 }, "evidence": [ { "supports": "In 2021, Yoseline Angel Lopez's team's winning web-app prototype was adopted by Colombia's Ministry of Agriculture, connecting smallholder farmers with local buyers; also supported agricultural land-use monitoring using satellite imagery", "url": "https://www.nasa.gov/directorates/stmd/prizes-challenges-crowdsourcing-program/registration-opens-for-2025-nasa-international-space-apps-challenge/" } ], "hot": 3, "id": "nasa-space-apps-colombia", "links": { "official": "https://spaceappschallenge.org" }, "name": "NASA Space Apps Colombia agriculture platform", "tags": [ "agriculture", "satellite", "open-data" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nNeutun Labs built a seizure monitoring platform on a smartwatch. The watch detects seizure-like activity and logs it for the wearer and their clinician.\n\n## The problem\n\nSeizures are often unwitnessed. A patient's own report is unreliable, and a diary misses events that happen during sleep.\n\n## What makes it interesting\n\nIt won the Pebble prize for best smartwatch app at Hack the North 2014 and continued development after the event. The project used a consumer watch rather than custom medical hardware.\n\n## What you can learn\n\nA general-purpose device can be a clinical sensor if the signal is good enough. That trade changes the cost and the adoption path.\n\n## Limitations of this analysis\n\nThe record does not describe the detection algorithm, its sensitivity, or any regulatory clearance.\n", "description": "Seizure monitoring smartwatch platform. Won the Pebble prize for best smartwatch app and continued development after the hackathon.", "domain": "healthcare", "event": { "award": "Pebble prize for best smartwatch app", "name": "Hack the North", "year": 2014 }, "evidence": [ { "supports": "Hack the North 2014 Pebble prize win", "url": "https://betakit.com/collaboration-trumps-competition-at-hackthenorth/" } ], "hot": 20, "id": "neutun-labs", "name": "Neutun Labs", "tags": [ "wearable", "health", "college-hackathon" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nNo Humanity is a multiplayer survival game built to be extremely difficult. It has drawn more than 20 million players across platforms.\n\n## What makes it interesting\n\nIt started at a Global Game Jam. Difficulty became the marketing: a game that most players cannot finish produces clips, complaints, and streams.\n\n## What you can learn\n\nFrustration is a shareable emotion. If the failure is fair and funny, players will advertise the game for you.\n\n## Limitations of this analysis\n\nThe record does not describe the mechanics, the platforms, or how retention held after the initial spike.\n", "description": "Extremely difficult multiplayer survival game. Over 20 million players across platforms.", "domain": "games", "event": { "name": "Global Game Jam" }, "evidence": [ { "supports": "GGJ origin and player count", "url": "https://globalgamejam.org/ggj-success-stories" } ], "hot": 10, "id": "no-humanity", "links": { "official": "https://nohumanitygame.com/" }, "name": "No Humanity", "notes": "Origin at Global Game Jam, developed by SweatyChair Studio", "tags": [ "game-jam", "multiplayer", "indie" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nNoonlight is a personal safety service. A user triggers an alarm from a phone or a connected device, and a dispatch center responds.\n\n## The problem\n\nCalling emergency services requires speaking, knowing where you are, and having time. In a dangerous situation, none of those may be available.\n\n## How the experience works\n\nThe user raises an alarm from the app or a hardware button. The service contacts emergency responders with the user's location and details, without the user having to hold a call.\n\n## What makes it interesting\n\nIt began as a Startup Weekend project in 2013 and became a service integrated into dating apps and wearables. The integrations matter: safety is more useful when it sits inside the app where the risk happens.\n\n## What you can learn\n\nA safety feature that requires opening a separate app competes with panic. Putting it in the surface people already use removes that step.\n\n## Limitations of this analysis\n\nThe record does not describe the dispatch integration, false-alarm handling, or the hardware devices.\n", "description": "Personal safety platform that began as a Startup Weekend project; became a widely integrated panic-button service used by dating apps and smart devices.", "domain": "social", "event": { "award": "project origin", "name": "Startup Weekend", "organizer": "Startup Weekend", "year": 2013 }, "evidence": [ { "supports": "Noonlight profiled as a Kansas City startup that originated from a Startup Weekend project", "url": "https://www.startlandnews.com/2018/11/kansas-city-startup-noonlight-safety-app/" } ], "hot": 30, "id": "noonlight", "links": { "official": "https://www.noonlight.com" }, "name": "Noonlight", "tags": [ "startup-weekend", "safety", "mobile", "social-good" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nODIN is a marketplace for smart contract circuit breakers, bonded through EigenLayer. Protocols can buy protection that triggers when a contract misbehaves.\n\n## The problem\n\nA bug in a deployed contract can drain funds before a team can react. An automatic stop is only credible if someone has money at stake on the trigger being correct.\n\n## What makes it interesting\n\nIt won Best of ETHDenver 2024. Bonding the breaker to an economic stake is the design idea: the protection is only as good as the collateral behind it.\n\n## What you can learn\n\nSecurity claims need a cost for being wrong. A guarantee with no skin in the game is a promise.\n\n## Limitations of this analysis\n\nThe record does not describe the trigger conditions, the collateral terms, or whether any protocol used it.\n", "description": "Smart contract circuit breaker marketplace bonded to EigenLayer for on-chain security.", "domain": "fintech", "event": { "award": "Best of ETHDenver", "name": "ETHDenver", "year": 2024 }, "evidence": [ { "supports": "ETHDenver 2024 Best of ETHDenver award", "url": "https://unchainedcrypto.com/ethdenvers-2024-hackathon-onchain-security-ai-and-gaming-fun-dominate-the-winning-projects/" } ], "hot": 25, "id": "odin", "name": "ODIN", "tags": [ "web3", "security", "ethereum", "eigenlayer" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nOfficially Dead is a fast zombie-survival action game made for Ludum Dare 59 in 2025. It is being expanded into a Steam release by Erebus Games.\n\n## What makes it interesting\n\nIt ranked in the top 100 for humour and fun at the jam. A jam entry that is funny is easier to market than one that is merely polished.\n\n## What you can learn\n\nTone is a differentiator in a crowded genre. A joke that lands in the first minute is worth more than a feature list.\n\n## Limitations of this analysis\n\nThe record does not describe the combat or how the full release differs from the jam build.\n", "description": "Fast-paced zombie-survival action game from Ludum Dare 59; being expanded into a full commercial Steam release by Erebus Games.", "domain": "games", "event": { "award": "jam prototype", "name": "Ludum Dare 59", "organizer": "Ludum Dare", "year": 2025 }, "evidence": [ { "supports": "Officially Dead from LD59 is being turned into a full Steam game by the developers; ranked top 100 for humour and fun at the jam", "url": "https://ldjam.com/users/technicalitycreations/feed" } ], "hot": 35, "id": "officially-dead", "links": { "official": "https://ldjam.com/events/ludum-dare/59/officially-dead" }, "name": "Officially Dead", "tags": [ "ludum-dare", "zombie", "action", "steam" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nOpenAI Gym is a standard interface for reinforcement learning environments. An agent sees an observation, picks an action, and receives a reward and a new observation. Gym defines the shape of that exchange so the same training loop can drive a cart-pole simulation, an Atari game, or a robot arm.\n\n## The key insight\n\nBefore Gym, most papers shipped their own environment code and their own training harness. Results were hard to compare because the loop itself differed between implementations. Gym standardized the environment side of the loop and left the algorithm side alone. That is a much smaller surface than a full framework, and it was enough.\n\n## How it works\n\nEvery environment implements the same contract:\n\n- `reset()` returns the first observation.\n- `step(action)` returns observation, reward, done, and an info dictionary.\n- `action_space` and `observation_space` describe legal actions and observations.\n\nSpaces are typed. `Discrete(4)` is a set of four actions. `Box(low, high, shape)` is a continuous array with bounds. Because the shapes are declared, an algorithm can build the right input and output layers without knowing which environment it is talking to.\n\nWrappers modify an environment while keeping the contract. Frame stacking, reward scaling, and time limits are all wrappers: they take an environment and return an environment. This is composition instead of configuration, and it is why third-party modifications do not require changes to the core.\n\nRegistration maps a string like `CartPole-v1` to a constructor. `gym.make` builds the environment and wraps it. Versioning in the name matters: `v0` and `v1` can differ in reward definition or episode length, and published results cite the version.\n\n## Build something similar\n\nDefine the contract first, in fewer than ten lines. Two methods and two space objects are enough.\n\nBuild one environment with a trivial dynamic, then write a random agent that samples from the action space and prints the episode return. If the agent can run without importing anything specific to that environment, the interface works.\n\nThen add a wrapper that changes one thing, such as stacking the last four observations. If the wrapper needs no change to the agent, the composition boundary is in the right place.\n\n## Sources\n\nThe interface and repository are linked in the evidence field.\n", "deep_dive": true, "description": "Toolkit for reinforcement learning research; described by OpenAI as not a formal hackathon project but built with rapid-iteration hackathon culture in mind.", "domain": "ai-ml", "evidence": [ { "supports": "OpenAI Gym repository and docs; OpenAI has stated it was not a formal hackathon project but embodied rapid-build research culture", "url": "https://github.com/openai/gym" } ], "hot": 35, "id": "openai-gym", "key_idea": "Give every environment the same small interface, so one algorithm runs everywhere without adaptation.", "learn_from": "A boring, fixed contract creates an ecosystem. Five methods beat a clever framework.", "links": { "official": "https://github.com/openai/gym", "source": "https://github.com/openai/gym" }, "name": "OpenAI Gym", "tags": [ "reinforcement-learning", "open-source", "research", "rapid-build" ], "verified_at": "2026-10-24" }, { "authors": [ "Scott Fitsimones", "Nik Shevchenko", "Stepan Korshakov", "Arash Ramanan", "Jatin Gupta" ], "body": "## What it is\n\nOpenGlass is an open-source attachment that turns ordinary glasses into smart glasses. A camera and a small computer clip onto the frame, and a model describes what the camera sees.\n\n## The problem\n\nSmart glasses were expensive and closed. Developers who wanted to experiment with wearable vision had to buy a finished product and work inside its limits.\n\n## How the experience works\n\nThe hardware mounts on existing frames. The camera captures a scene, and a vision model produces a description that is spoken or displayed to the wearer.\n\n## What makes it interesting\n\nA five-person team built it in about 36 hours at the Cerebral Valley hackathon in 2024 for roughly $20 in off-the-shelf parts, and won the event. The bill of materials is the point: a wearable anyone can rebuild.\n\n## What you can learn\n\nOpen hardware wins attention when the parts list is short and the result is recognizable. A $20 demo does more for adoption than a specification.\n\n## Limitations of this analysis\n\nThe record does not document the compute module, battery life, latency, or the split between on-device and remote inference.\n", "description": "Open-source AI smart-glasses attachment built in about 36 hours for roughly $20 of off-the-shelf parts, pairing a camera with Llama 3 for live scene description.", "domain": "hardware", "event": { "award": "hackathon winner", "name": "Cerebral Valley Hackathon", "organizer": "Cerebral Valley", "year": 2024 }, "evidence": [ { "supports": "Five-person team built $20 open-source smart glasses at Cerebral Valley hackathon in a weekend; won the event and collected roughly 1,500 waitlist signups", "url": "https://gizmodo.com/cerebral-valley-hackers-build-20-open-source-smart-gla-1851476656" }, { "supports": "36-hour hackathon project described as $20 open-source smart glasses connecting sight and hearing to an AI chatbot", "url": "https://qz.com/a-group-of-hackers-built-20-open-source-smart-glasses-1851478685" }, { "supports": "Open-source repo: turn any glasses into hackable smart glasses with under $25 of off-the-shelf components", "url": "https://github.com/not-important-vr/openglass" } ], "hot": 65, "id": "openglass", "links": { "demo": "https://github.com/not-important-vr/openglass", "official": "https://github.com/BurgerKing6668/OpenGlass2", "source": "https://github.com/BurgerKing6668/OpenGlass2" }, "name": "OpenGlass", "tags": [ "ai-wearables", "open-source", "accessibility", "hackathon", "playable" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nOpenVent is an open-source ventilator design produced during Infineon Technologies' COVID-era hackathon in 2020. It uses an Ambu bag as the air source.\n\n## The problem\n\nVentilators were scarce in early 2020. A design that can be built locally from common parts is a different response from waiting for a factory.\n\n## What makes it interesting\n\nIt is one of several rapid-build emergency respiratory projects from that period. The Ambu-bag approach avoids designing a compressor and focuses on control and sensing.\n\n## What you can learn\n\nIn an emergency, the supply chain is part of the design. A device built from available parts can matter more than a better device that cannot be manufactured.\n\n## Limitations of this analysis\n\nThe record does not include test results, regulatory status, or whether the design was built and used. Open-source ventilation is a safety-critical claim.\n", "description": "Open-source ventilator design produced during Infineon Technologies' COVID-era hackathon; one of several rapid-build emergency respiratory device projects in 2020.", "domain": "healthcare", "event": { "award": "hackathon project", "name": "Infineon Technologies Hackathon", "organizer": "Infineon Technologies", "year": 2020 }, "evidence": [ { "supports": "OpenVent developed an open-source ventilator with an Ambu-bag during Infineon Technologies' COVID-19 hackathon", "url": "https://www.medicaldesignandoutsourcing.com/can-an-open-source-ventilator-hackathon-help-fight-the-coronavirus" } ], "hot": 33, "id": "openvent", "links": { "official": "https://www.medicaldesignandoutsourcing.com/can-an-open-source-ventilator-hackathon-help-fight-the-coronavirus" }, "name": "OpenVent", "tags": [ "open-hardware", "covid-19", "ventilator", "hackathon" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nPathSense is an indoor navigation app for blind and low-vision users. It provides spoken guidance through buildings where GPS does not work.\n\n## The problem\n\nOutdoor navigation is solved by satellite positioning. Indoors, a user needs a map and a position, and neither is usually available.\n\n## What makes it interesting\n\nIt was a finalist at Hack the North 2024. The project targets the last few meters, which is where wayfinding actually fails.\n\n## What you can learn\n\nIndoor positioning is an infrastructure problem before it is an app problem. Someone has to map the building.\n\n## Limitations of this analysis\n\nThe record does not describe the positioning method or any user testing.\n", "description": "AI indoor navigation for vision-impaired users with voice guidance.", "domain": "accessibility", "event": { "name": "Hack the North", "year": 2024 }, "evidence": [ { "supports": "Hack the North 2024 finalist", "url": "https://museum.hackthenorth.com/" } ], "hot": 5, "id": "pathsense", "name": "PathSense", "tags": [ "accessibility", "ai-ml", "navigation", "college-hackathon" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nPDFi turns static PDFs into interactive financial documents. Payments and transactions can be embedded in the file itself.\n\n## The problem\n\nA financial document, such as an invoice or a term sheet, is a dead file. Acting on it means leaving the document for a banking app.\n\n## What makes it interesting\n\nIt won Best Finance App on Fern at ETHGlobal New York 2025. The PDF is a format every business already uses, which is why the demo lands.\n\n## What you can learn\n\nAttaching an action to a familiar artifact is a strong wedge. The user does not have to adopt a new workflow.\n\n## Limitations of this analysis\n\nThe record does not describe the signing model, the chain integration, or the security of embedding transactions in a file.\n", "description": "Turns static PDFs into live financial tools with embedded cryptocurrency transactions; won Best Finance App at ETHGlobal New York 2025.", "domain": "fintech", "event": { "award": "Best Finance App on Fern", "name": "ETHGlobal New York", "organizer": "ETHGlobal", "year": 2025 }, "evidence": [ { "supports": "PDFi won Best Finance App on Fern at ETHGlobal New York 2025 (August 15-17); reimagined PDFs as live financial tools with crypto transactions embedded in the document", "url": "https://fernhq.com/blog/ethglobal-new-york-2025-fern-api-prize-winners" } ], "hot": 25, "id": "pdfi", "links": { "official": "https://ethglobal.com/showcase/pdfi-z99gp" }, "name": "PDFi", "tags": [ "ethglobal", "pdf", "crypto", "finance" ], "verified_at": "2026-10-24" }, { "authors": [ "Aggro Crab", "Landfall Games" ], "body": "## What it is\n\nPeak is a cooperative climbing game with physics-driven movement. Players climb a mountain together and need each other to get up.\n\n## What makes it interesting\n\nAggro Crab and Landfall jammed it in early 2025 and shipped it months later. It sold over 2 million copies on Steam.\n\n## What you can learn\n\nA co-op game spreads through pairs of players. The purchase is social, so the game has to be funny before it is good.\n\n## Limitations of this analysis\n\nThe record does not describe the climbing systems or how the jam version changed before release.\n", "description": "Physics co-op climbing game jammed in early 2025 by Aggro Crab and Landfall; shipped on Steam months later and sold over 2 million copies.", "domain": "games", "event": { "award": "jam prototype", "name": "Peak Jam", "organizer": "Aggro Crab and Landfall", "year": 2025 }, "evidence": [ { "supports": "Peak sold over 2 million copies on Steam after a rapid jam-to-release development cycle in 2025", "url": "https://www.pcgamer.com/games/co-op-physics-climbing-game-peak-has-already-sold-2-million-copies/" } ], "hot": 35, "id": "peak", "links": { "official": "https://store.steampowered.com/app/2527290/PEAK/" }, "name": "Peak", "tags": [ "game-jam", "co-op", "indie", "steam" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nPlease, Don't Touch Anything is a puzzle game about a console with one button and a note telling you not to press it. Each press changes the room and the puzzle.\n\n## What makes it interesting\n\nFour Quarters developed it after a jam prototype. The premise is a single forbidden action, which is enough to carry a short game and a lot of curiosity.\n\n## What you can learn\n\nA prohibition is a prompt. Telling the player not to do something is often more effective than telling them what to do.\n\n## Limitations of this analysis\n\nThe record does not describe the puzzle tree or the number of endings.\n", "description": "Puzzle box game about a mysterious button and a forbidden console. Developed by Four Quarters after a jam prototype.", "domain": "games", "event": { "name": "Ludum Dare" }, "evidence": [ { "supports": "Four Quarters jam origin", "url": "https://www.vice.com/en/article/how-the-loop-hero-devs-pitched-a-hit-game-thats-impossible-to-describe/" } ], "hot": 10, "id": "please-dont-touch-anything", "links": { "official": "https://store.steampowered.com/app/370310/Please_Dont_Touch_Anything/" }, "name": "Please, Don't Touch Anything", "tags": [ "game-jam", "puzzle", "indie" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nPointshop Space is an augmented-reality shopping prototype. It places products in the user's space so they can be viewed before purchase.\n\n## The problem\n\nOnline shopping hides scale. A photo does not tell you whether a chair fits the room or how a jacket hangs.\n\n## What makes it interesting\n\nIt won the grand prize at the TechCrunch Disrupt San Francisco 2016 Hackathon. AR was new on phones that year, and the demo used the device's camera rather than a headset.\n\n## What you can learn\n\nA demo that changes what the judge sees in the room beats a slide. The stage becomes the product surface.\n\n## Limitations of this analysis\n\nThe record does not describe the tracking method or any retail partnership.\n", "description": "Augmented-reality shopping prototype that won the TechCrunch Disrupt San Francisco 2016 Hackathon grand prize.", "domain": "web-mobile", "event": { "award": "grand prize", "name": "TechCrunch Disrupt San Francisco Hackathon", "organizer": "TechCrunch", "year": 2016 }, "evidence": [ { "supports": "Pointshop Space named grand prize winner of the TechCrunch Disrupt SF 2016 Hackathon", "url": "https://techcrunch.com/2016/09/13/pointshop-space-wins-disrupt-sf-2016-hackathon/" } ], "hot": 31, "id": "pointshop-space", "links": { "official": "https://techcrunch.com/2016/09/13/pointshop-space-wins-disrupt-sf-2016-hackathon/" }, "name": "Pointshop Space", "tags": [ "techcrunch-disrupt", "ar", "retail" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nPony Island is a puzzle game disguised as a broken arcade machine. The machine is run by the devil, and the game keeps breaking its own rules.\n\n## What makes it interesting\n\nDaniel Mullins developed it primarily during Ludum Dare 31 in December 2014, in 48 hours, on the theme \"Entire Game on One Screen\". It later became a Steam release.\n\n## What you can learn\n\nThe metafictional trick only works if the underlying game is playable. The joke is layered on a real loop, not instead of one.\n\n## Limitations of this analysis\n\nThe record does not describe the puzzle design or how the commercial version expanded the jam.\n", "description": "Metafictional puzzle game disguised as a broken arcade machine designed by the devil.", "domain": "games", "event": { "name": "Ludum Dare 31", "year": 2014 }, "evidence": [ { "supports": "Primarily developed during the 48-hour Ludum Dare game jam in December 2014, theme 'Entire Game On One Screen'", "url": "https://en.wikipedia.org/wiki/Pony_Island" }, { "supports": "Develop post-mortem with Daniel Mullins describing the Ludum Dare prototype becoming a Steam release", "url": "https://mcvuk.com/development-news/the-develop-post-mortem-pony-island" }, { "supports": "Original jam version page: 'developed for Ludum Dare 31 in 48 hours'", "url": "https://dmullinsgames.itch.io/pony-island" } ], "hot": 10, "id": "pony-island", "links": { "official": "https://store.steampowered.com/app/405640/Pony_Island/" }, "name": "Pony Island", "tags": [ "puzzle", "metafiction", "horror" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nSession Replay records a user's clicks, scrolls, and page transitions so a product team can watch what happened without reproducing it.\n\n## The problem\n\nAnalytics showed that users dropped off; it did not show why. Watching a real session closes the gap between a funnel chart and a bug report.\n\n## What makes it interesting\n\nAn engineer built it during a PostHog hackathon in the company's early days, against leadership's scepticism. It grew to tens of thousands of daily active users and became one of the products that turned PostHog from an analytics tool into a multi-product platform.\n\n## What you can learn\n\nA hackathon prototype can settle a strategy argument that meetings cannot. The cost of being wrong is a weekend; the cost of debating is a quarter.\n\n## Limitations of this analysis\n\nThe record does not describe the recording format, the privacy controls, or how replay data is stored.\n", "description": "Product analytics session-replay feature built during an early PostHog internal hackathon; grew into a core multi-product pillar with tens of thousands of daily active users.", "domain": "developer-tools", "event": { "name": "PostHog Internal Hackathon", "organizer": "PostHog" }, "evidence": [ { "supports": "In PostHog's earliest days an engineer built Session Replay during a hackathon despite leadership pushback; it became wildly popular with 50,000+ DAU and drove the multi-product strategy", "url": "https://newsletter.posthog.com/p/great-companies-are-built-in-hackathons" }, { "supports": "PostHog states session replay and the data warehouse started as hackathon projects", "url": "https://posthog.com/blog/behind-the-scenes-posthog-hackathon" } ], "hot": 10, "id": "posthog-session-replay", "links": { "official": "https://posthog.com/session-replay" }, "name": "PostHog Session Replay", "tags": [ "analytics", "internal-hackathon", "product-analytics" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nProject M is an emotional AI system that reads micro-facial expressions and reports an emotional state. It targets clinical use, including pediatric mental telehealth.\n\n## The problem\n\nEmotional assessment in telehealth depends on self-report and clinician observation, both of which lose information over a video call.\n\n## What makes it interesting\n\nIt won the Touchless Experience category at the Texas Children's Hospital Healthcare Hackathon in 2021. The team planned FDA Software as a Medical Device registration, a much longer road than a hackathon.\n\n## What you can learn\n\nA clinical demo is the start of a regulatory process, not the end of development. The decision to pursue clearance is the part worth studying.\n\n## Limitations of this analysis\n\nThe record does not include accuracy results, training data, or current regulatory status. Claims about clinical validity cannot be checked from the linked source.\n", "description": "Emotional AI platform detecting micro-facial expressions for clinical use; won the Texas Children's Hospital Healthcare Hackathon Touchless Experience category in 2021.", "domain": "healthcare", "event": { "award": "Touchless Experience category winner", "name": "Texas Children's Hospital Healthcare Hackathon", "organizer": "Texas Children's Hospital", "year": 2021 }, "evidence": [ { "supports": "IPMD's Project M won the Touchless Experience category at Texas Children's Hospital Healthcare Hackathon 2021; plans FDA Software as Medical Device registration for mental telehealth", "url": "https://developer.nvidia.com/blog/emotional-ai-solution-won-texas-childrens-hospital-healthcare-hackathon" } ], "hot": 33, "id": "project-m", "links": { "official": "https://developer.nvidia.com/blog/emotional-ai-solution-won-texas-childrens-hospital-healthcare-hackathon" }, "name": "Project M", "tags": [ "healthcare-hackathon", "ai", "emotion-recognition", "medical-device" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nProject Naptha is a browser extension that makes text inside images selectable, copyable, editable, and translatable. You drag over a screenshot or a meme as if it were a document, and the words come out.\n\n## The key insight\n\nOptical character recognition was not new in 2013. The insight was placement. Once the engine returns words with bounding boxes, the extension can position transparent text elements directly over the image. From that point, the browser's own selection, copy, and translation features do the rest.\n\nThe interface is the trick. Users do not learn a new tool. They use the selection behavior they already have.\n\n## How it works\n\nThe extension captures the image, draws it to a canvas, and runs an OCR engine compiled to JavaScript. Running locally avoids uploading every image to a server, which matters for privacy and for latency.\n\nThe engine returns recognized words and their coordinates. The extension creates a layer above the image with one positioned element per word or line. The text is transparent, but it sits in the same place as the pixels, so a drag selects the words under the cursor.\n\nBecause the layer is real text, other features follow. Copy puts the words on the clipboard. A translation action sends the selected string to a translation service. An edit action hides the original region and draws a replacement.\n\nDetection is the weak point. OCR on low-resolution or stylized text produces garbage, so the extension scores confidence and only builds overlays where the result is plausible. Images without text get no layer.\n\n## Build something similar\n\nTake one screenshot with clear text. Run any OCR engine over it and print each word with its bounding box.\n\nDraw absolutely positioned spans over the image using those boxes, with transparent color. Select them with the mouse. If selection lands on the right words, the core mechanism works.\n\nThen handle failure. Rotate the image, blur it, and use a decorative font. Decide what confidence threshold means \"no text here\" rather than producing nonsense.\n\n## Sources\n\nHackMIT origin is linked in the evidence field.\n", "deep_dive": true, "description": "Browser extension that lets users highlight, copy, edit, and translate text inside images.", "domain": "ai-ml", "event": { "name": "HackMIT", "year": 2013 }, "evidence": [ { "supports": "HackMIT 2013 origin", "url": "https://en.wikipedia.org/wiki/HackMIT" } ], "hot": 20, "id": "project-naptha", "key_idea": "Run OCR locally, then lay invisible text over the image so the recognized words behave like real text.", "learn_from": "The model is not the product. The product is making the output act like the interface people already know.", "links": { "official": "https://www.projectnaptha.com/" }, "name": "Project Naptha", "tags": [ "browser-extension", "ocr", "college-hackathon" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nProject Prometheus is a low-cost ventilator and respiratory support design built for the NASA Space Apps 2020 COVID-19 challenge. It was selected as the overall global winner.\n\n## The problem\n\nThe pandemic created a shortage of respiratory equipment, especially in places that could not buy from the usual manufacturers.\n\n## What makes it interesting\n\nThe project is a design response, not a software demo. Space Apps normally produces data tools, and this entry used the event to address a hardware crisis.\n\n## What you can learn\n\nA hackathon can coordinate a design effort, but it cannot certify a medical device. The value is in the open design and the people who pick it up.\n\n## Limitations of this analysis\n\nThe record does not report testing, regulatory review, or deployment. Safety-critical claims need evidence beyond a competition result.\n", "description": "COVID-era NASA Space Apps global winner building low-cost open ventilator and respiratory support designs for crisis conditions.", "domain": "healthcare", "event": { "award": "global winner", "name": "NASA Space Apps Challenge", "organizer": "NASA", "year": 2020 }, "evidence": [ { "supports": "Project Prometheus page for NASA Space Apps 2020 COVID-19 global challenge; selected as overall global winner", "url": "https://www.spaceappschallenge.org/2020/challenge/project-prometheus/" } ], "hot": 23, "id": "project-prometheus", "links": { "official": "https://www.spaceappschallenge.org/2020/challenge/project-prometheus/" }, "name": "Project Prometheus", "tags": [ "nasa-space-apps", "covid-19", "open-hardware", "medical" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nProof of Picture detects AI-generated images using C2PA provenance data. It is aimed at newsrooms checking images before publication.\n\n## The problem\n\nA generated image can look like a photograph. A newsroom needs a way to check origin without trusting the file's own metadata.\n\n## What makes it interesting\n\nIt won the AI and C2PA category at ETH Warsaw 2023. The project leans on a standard for content credentials rather than a classifier alone, which is a more durable bet.\n\n## What you can learn\n\nProvenance is only useful if the camera or editor signs at capture. Detection without adoption is a partial fix.\n\n## Limitations of this analysis\n\nThe record does not describe the verification flow or how it handles images with no credentials.\n", "description": "C2PA-based detection of AI-generated images to combat news misinformation.", "domain": "social", "event": { "award": "AI/C2PA winner", "name": "ETH Warsaw", "year": 2023 }, "evidence": [ { "supports": "ETH Warsaw 2023 AI/C2PA winner", "url": "https://www.recursive.so/hackathons" } ], "hot": 23, "id": "proof-of-picture", "name": "Proof of Picture", "tags": [ "web3", "ai-ml", "misinformation", "c2pa" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nPulseGrip is a motorized hand assist for people with paralysis. It reads muscle signals and closes the hand.\n\n## The problem\n\nGrip is one of the first functions lost after a spinal injury or nerve damage, and it is needed for almost every daily task.\n\n## What makes it interesting\n\nIt was a finalist at Hack the North 2024. The device uses EMG, so the user's remaining muscle signal becomes the control input.\n\n## What you can learn\n\nAssistive hardware should use the signal the body still produces. Asking for a new input method adds friction to every use.\n\n## Limitations of this analysis\n\nThe record does not describe the grip force, the fit, or any user trial.\n", "description": "EMG-powered hand-closing motor assistive device for people with paralysis.", "domain": "accessibility", "event": { "name": "Hack the North", "year": 2024 }, "evidence": [ { "supports": "Hack the North 2024 finalist", "url": "https://museum.hackthenorth.com/" } ], "hot": 5, "id": "pulsegrip", "name": "PulseGrip", "tags": [ "accessibility", "emg", "hardware", "college-hackathon" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nPush Protocol is a decentralized notification layer for Web3 applications. It lets a wallet or app receive messages without polling a chain.\n\n## The problem\n\nOn-chain applications have no native way to reach a user. A transaction confirmation or a liquidation warning has to be discovered by refreshing.\n\n## What makes it interesting\n\nIt started at ETHGlobal 2020 and later raised $10 million. Notifications are unglamorous infrastructure, which is why they are a durable position once apps integrate.\n\n## What you can learn\n\nEvery application stack needs a messaging layer. Owning the default channel is more valuable than owning one app.\n\n## Limitations of this analysis\n\nThe record does not describe the protocol design or how spam and cost are handled.\n", "description": "Decentralized push notification protocol for Web3 applications. Raised $10 million in funding.", "domain": "web-mobile", "event": { "name": "ETHGlobal", "year": 2020 }, "evidence": [ { "supports": "ETHGlobal 2020 origin and funding round", "url": "https://bizthon.hashnode.dev/from-hackathon-to-unicorn-real-stories-of-web3-projects-that-made-it-big" } ], "hot": 3, "id": "push-protocol", "links": { "official": "https://push.org/" }, "name": "Push Protocol", "tags": [ "web3", "notifications", "ethereum" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nQuick Insurance is a mobile prototype for buying short-term coverage on a valuable item with a few taps.\n\n## The problem\n\nInsurance is sold in annual policies with forms and agents. A person who needs cover for an afternoon, a bicycle, or a camera has no easy option.\n\n## What makes it interesting\n\nIt won the Grand Prize and €4,200 at the TechCrunch Disrupt Berlin 2017 Hackathon. The demo was the purchase flow, which is the part incumbents make painful.\n\n## What you can learn\n\nIn insurance, distribution and paperwork are the product. Shortening the flow is the innovation.\n\n## Limitations of this analysis\n\nThe record does not describe underwriting, claims, or whether the product launched.\n", "description": "Tap-to-insure mobile prototype for short-term valuable-item coverage; won the TechCrunch Disrupt Berlin 2017 Hackathon Grand Prize and €4,200.", "domain": "fintech", "event": { "award": "Grand Prize", "name": "TechCrunch Disrupt Berlin Hackathon", "organizer": "TechCrunch", "year": 2017 }, "evidence": [ { "supports": "Quick Insurance won the Disrupt Berlin 2017 Hackathon Grand Prize and €4,200 for a few-taps insurance purchase flow", "url": "https://techcrunch.com/2017/12/03/quick-insurance-wins-the-disrupt-berlin-2017-hackathon-grand-prize" } ], "hot": 31, "id": "quick-insurance", "links": { "official": "https://techcrunch.com/2017/12/03/quick-insurance-wins-the-disrupt-berlin-2017-hackathon-grand-prize" }, "name": "Quick Insurance", "tags": [ "techcrunch-disrupt", "insurance", "mobile", "hackathon-winner" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nRainmaker: The Beautiful Flood is a puzzle game. It started at Global Game Jam 2014 and later reached the App Store, where it was featured by Apple.\n\n## What makes it interesting\n\nThe jam entry became a mobile release with an Apple feature. A puzzle game is a good fit for a phone, and the jam scope kept the design small.\n\n## What you can learn\n\nA feature placement is a distribution event. Designing for a platform's editorial taste can matter as much as the mechanic.\n\n## Limitations of this analysis\n\nThe record does not describe the puzzle mechanics or the game's performance after launch.\n", "description": "Puzzle game featuring Apple Best New Game global feature and", "domain": "games", "event": { "name": "Global Game Jam 2014" }, "evidence": [ { "supports": "GGJ 2014 origin, Apple feature, and App Store rankings", "url": "https://www.richardfu.net/dev-story-behind-rainmaker-the-beautiful-flood/" } ], "hot": 10, "id": "rainmaker", "links": { "official": "https://www.richardfu.net/rainmaker/" }, "name": "Rainmaker: The Beautiful Flood", "tags": [ "game-jam", "puzzle", "mobile" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nResonant Exoplanets is an AI system for detecting exoplanets. It looks for the signals that a planet leaves in observations of its star.\n\n## What makes it interesting\n\nIt won Best Use of Technology at NASA Space Apps 2025. Detection is a needle-in-haystack problem, which is why machine learning fits the data.\n\n## What you can learn\n\nIn discovery work, the cost of a false positive is an astronomer's time. A model should rank candidates, not declare planets.\n\n## Limitations of this analysis\n\nThe record does not describe the data, the model, or its precision.\n", "description": "AI-based exoplanet detection system.", "domain": "science", "event": { "award": "Best Use of Technology", "name": "NASA Space Apps Challenge", "year": 2025 }, "evidence": [ { "supports": "Space Apps 2025 Best Use of Technology award", "url": "https://www.spaceappschallenge.org/2025/awards/global-winners/" } ], "hot": 25, "id": "resonant-exoplanets", "name": "Resonant Exoplanets", "tags": [ "astronomy", "ai-ml", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nRollerdrome is a third-person shooter that mixes rollerskating with gunplay. The player performs tricks and fights in an arena sport.\n\n## How the experience works\n\nMovement is the defense: skating fast and tricking keeps the player alive while aiming. The arena is built for loops and ramps rather than cover.\n\n## What makes it interesting\n\nPaul Rabbitte conceived it during a 48-hour Game Maker's Toolkit jam in 2017, and it later became a full release. The jam version proved that skating and shooting could share one control scheme.\n\n## What you can learn\n\nTwo verbs that normally conflict can be the hook. If the prototype makes the conflict fun, the design is already there.\n\n## Limitations of this analysis\n\nThe record does not describe the combat tuning or how the jam concept became a commercial game.\n", "description": "Third-person action-shooter that blends rollerskating tricks with gunplay in a dystopian retrofuture arena sport.", "domain": "games", "event": { "name": "GMTK Game Jam", "year": 2017 }, "evidence": [ { "supports": "Initially conceived by Paul Rabbitte during a 48-hour game jam hosted by Game Maker's Toolkit in 2017", "url": "https://en.wikipedia.org/wiki/Rollerdrome" } ], "hot": 11, "id": "rollerdrome", "links": { "official": "https://rollerdromegame.com" }, "name": "Rollerdrome", "tags": [ "action", "shooter", "skating" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nThe Roottrees are Dead is a deduction game. The player researches a 1998 candy-company inheritance and fills in a family tree.\n\n## How the experience works\n\nClues arrive through documents, photos, and records. The player cross-references names and dates, then places each person in the tree.\n\n## What makes it interesting\n\nJeremy Johnston built it for Global Game Jam 2023 in under a week, released it as a free browser game, and later rebuilt it with hand-drawn art for Steam.\n\n## What you can learn\n\nA deduction game needs a truth behind the puzzle. The solver should be able to prove an answer, not guess it.\n\n## Limitations of this analysis\n\nThe record does not describe the clue design or the changes made for the commercial release.\n", "description": "Deduction puzzle game where players research a 1998 candy-company inheritance to fill in an intricate family tree.", "domain": "games", "event": { "name": "Global Game Jam 2023", "year": 2023 }, "evidence": [ { "supports": "Created for a Global Game Jam by Jeremy Johnston in under a week in 2023", "url": "https://en.wikipedia.org/wiki/The_Roottrees_Are_Dead" }, { "supports": "Originally a Global Game Jam project cobbled together in less than a week in 2023; later rebuilt with hand-drawn art", "url": "https://www.theverge.com/ai-artificial-intelligence/686651/roottrees-ai-original-illustrator-replacement" }, { "supports": "Steam page states the game was originally created for Global Game Jam 2023 and released as a free browser game", "url": "https://store.steampowered.com/app/2754380/The_Roottrees_are_Dead/" } ], "hot": 13, "id": "roottrees-are-dead", "links": { "official": "https://store.steampowered.com/app/2754380/The_Roottrees_are_Dead/" }, "name": "The Roottrees are Dead", "tags": [ "puzzle", "detective", "research" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nROSS is an art access tool for blind users. It conveys visual content through touch, temperature, and sound.\n\n## The problem\n\nA museum audio guide describes a painting in words. That tells you what the guide noticed, not what is in front of you.\n\n## What makes it interesting\n\nIt was a finalist at Hack the North 2025. Using temperature as a channel is unusual and gives the user a sense that is not already occupied by audio.\n\n## What you can learn\n\nAccessibility design is often a mapping problem. The question is which senses are free and which properties fit them.\n\n## Limitations of this analysis\n\nThe record does not describe the hardware or any testing with blind users.\n", "description": "Art access tool for blind users that conveys visual content through touch, temperature, and sound.", "domain": "accessibility", "event": { "name": "Hack the North", "year": 2025 }, "evidence": [ { "supports": "Hack the North 2025 finalist", "url": "https://museum.hackthenorth.com/" } ], "hot": 5, "id": "ross-remote-operated-semantic-sketching", "name": "ROSS - Remote Operated Semantic Sketching", "tags": [ "accessibility", "art", "sensory", "college-hackathon" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nRubble Busters is an autonomous drone system for disaster search. It navigates rubble and reports what it finds.\n\n## The problem\n\nAfter a collapse, the first hours are spent searching spaces too dangerous or too narrow for people. A drone can enter without endangering a responder.\n\n## What makes it interesting\n\nThe team won first place at Kent State's 2023 SkyHack. The task is a full loop: navigate, detect, and report, not just fly.\n\n## What you can learn\n\nA search robot is judged on what it reports. Autonomy that finds nothing is a camera with extra steps.\n\n## Limitations of this analysis\n\nThe record does not describe the sensors, the flight time, or any field exercise.\n", "description": "Autonomous search-and-report drone system; won first place at Kent State's 2023 SkyHack for disaster rubble navigation.", "domain": "robotics", "event": { "award": "first place", "name": "SkyHack", "organizer": "Kent State University", "year": 2023 }, "evidence": [ { "supports": "Team Rubble Busters won first place at the 2023 SkyHack at Kent State with an autonomous search-and-report disaster drone", "url": "https://www.kent.edu/today/news/rubble-busters-skyhack-team-creates-lifesaving-drone-fly" } ], "hot": 23, "id": "rubble-busters", "links": { "official": "https://www.kent.edu/today/news/rubble-busters-skyhack-team-creates-lifesaving-drone-fly" }, "name": "Rubble Busters", "tags": [ "skyhack", "drone", "search-and-rescue", "university-hackathon" ], "verified_at": "2026-10-24" }, { "body": "Longer than a weekend, shorter than traditional development: eight weeks of jam production produced a first-week million-plus seller. A new data point for jam economics.\n", "description": "Cooperative RV road-trip game developed during an eight-week summer game jam; hit 1.2 million sales in its first week on Steam in October 2025.", "domain": "games", "event": { "award": "jam prototype", "name": "Eight-Week Summer Game Jam", "organizer": "Independent collective", "year": 2025 }, "evidence": [ { "supports": "RV There Yet, developed during an eight-week summer game jam, sold 1.2 million copies in its first week on Steam", "url": "https://www.pcgamer.com/games/co-op-rv-driving-sim-rv-there-yet-sold-12-million-copies-in-its-first-week-on-steam/" } ], "hot": 35, "id": "rv-there-yet", "links": { "official": "https://store.steampowered.com/app/2518530/RV_There_Yet/" }, "name": "RV There Yet", "tags": [ "game-jam", "co-op", "simulation", "steam" ], "verified_at": "2026-10-24" }, { "body": "Enterprise hackathons often stall in demo purgatory. Sabre's concierge AI prototype crossed the gap into real travel-agent tooling.\n", "description": "AI travel concierge prototype built during an internal Sabre hackathon; scaled into production-facing assistant tooling for travel agents.", "domain": "developer-tools", "event": { "name": "Sabre Internal Hackathon", "organizer": "Sabre" }, "evidence": [ { "supports": "Sabre documents internal hackathon projects including AI concierge prototypes that moved toward production", "url": "https://www.sabre.com/insights/sabre-hackathon-innovation-ai-travel/" } ], "hot": 10, "id": "sabre-concierge-ai", "links": { "official": "https://www.sabre.com" }, "name": "Sabre Concierge AI", "tags": [ "internal-hackathon", "ai", "travel", "enterprise" ], "verified_at": "2026-10-24" }, { "body": "Scintilla turned distributed air quality feeds into live maps, earning NASA Space Apps 2016's Best Use of Data title.\n", "description": "Real-time air quality visualization platform built at NASA Space Apps 2016; won Best Use of Data for turning open sensor feeds into actionable maps.", "domain": "climate", "event": { "award": "Best Use of Data", "name": "NASA Space Apps Challenge", "organizer": "NASA", "year": 2016 }, "evidence": [ { "supports": "Scintilla project page for NASA Space Apps 2016; won Best Use of Data for air quality visualization", "url": "https://www.spaceappschallenge.org/2016/challenge/scintilla/" } ], "hot": 21, "id": "scintilla", "links": { "official": "https://www.spaceappschallenge.org/2016/challenge/scintilla/" }, "name": "Scintilla", "tags": [ "nasa-space-apps", "air-quality", "data-visualization", "climate" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nScreencheat is a split-screen shooter where every player is invisible. You find opponents by looking at their quarter of the screen.\n\n## The problem\n\nSplit-screen shooters assume each player sees only their own view. Screen looking is treated as cheating, and games hide information to prevent it.\n\n## How the experience works\n\nFour players share one screen. Since no one is visible, the only way to locate a target is to read the other panels: map geometry, weapon fire, the position of the camera.\n\n## What makes it interesting\n\nIt was built at Global Game Jam 2014 in Melbourne, then released on Steam, PlayStation 4, and Xbox One. The design turns an unsolvable hardware constraint into the rule of the game. You cannot stop players from looking at each other's screens, so the game is built on it.\n\n## What you can learn\n\nWhen a limitation cannot be removed, make it the mechanic. The cheapest fix is often a rule change rather than a technical one.\n\n## Limitations of this analysis\n\nThe record does not describe the netcode or how the console versions handle four-player split-screen performance.\n", "description": "Split-screen first-person shooter where all players are invisible, so you must cheat by looking at their screens. Released on Steam, PS4, Xbox One.", "domain": "games", "event": { "name": "Global Game Jam 2014", "organizer": "Melbourne site" }, "evidence": [ { "supports": "GGJ 2014 Melbourne origin and console releases", "url": "https://globalgamejam.org/ggj-success-stories" } ], "hot": 10, "id": "screencheat", "links": { "official": "https://screencheatgame.com/" }, "name": "Screencheat", "tags": [ "game-jam", "local-multiplayer", "fps" ], "verified_at": "2026-10-24" }, { "body": "Built by Sebastian Krzyszkowiak, Konrad Burandt, and Paweł Radej in the jam under the name \"The Game Jam Game About Games, Secrets and Stuff.\" Shipped on Steam October 2015 with three expansions.\n", "description": "Free point-and-click adventure where a gamedev intern infiltrates a corporation that controls the entire gaming industry.", "domain": "games", "event": { "name": "Global Game Jam 2015", "year": 2015 }, "evidence": [ { "supports": "Originally created for Global Game Jam 2015 in Poznań, Poland; published on Steam at the end of the same year", "url": "https://globalgamejam.org/ggj-success-stories" }, { "supports": "Official site states the first version was made during Global Game Jam 2015 in Poznań", "url": "https://tremendouscorp.com" } ], "hot": 11, "id": "secret-of-tremendous-corporation", "links": { "official": "https://store.steampowered.com/app/380140/The_Secret_of_Tremendous_Corporation/" }, "name": "The Secret of Tremendous Corporation", "tags": [ "point-and-click", "adventure", "satire" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nSekai is a story-building platform that uses AI to help people create and remix interactive stories. It runs on Story Protocol, which handles licensing of the underlying intellectual property.\n\n## The problem\n\nA collaborative story has an ownership problem. If several people contribute characters and worlds, nobody can easily license what they made or get paid when someone else builds on it.\n\n## What makes it interesting\n\nIt won Best of ETHDenver 2024. The project pairs a consumer-facing creative tool with on-chain licensing, which is harder than either piece alone.\n\n## What you can learn\n\nCreative tools need a reason to be on a chain. Here the reason is attribution and licensing, not payments for their own sake.\n\n## Limitations of this analysis\n\nThe record does not describe the AI pipeline, the licensing terms, or how a contributor's share is calculated.\n", "description": "AI-interactive story-building platform built on Story Protocol for IP licensing.", "domain": "media-arts", "event": { "award": "Best of ETHDenver", "name": "ETHDenver", "year": 2024 }, "evidence": [ { "supports": "ETHDenver 2024 Best of ETHDenver award", "url": "https://unchainedcrypto.com/ethdenvers-2024-hackathon-onchain-security-ai-and-gaming-fun-dominate-the-winning-projects/" } ], "hot": 25, "id": "sekai", "links": { "official": "https://sekai.co/" }, "name": "Sekai", "tags": [ "web3", "ai-ml", "storytelling", "ethereum" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nSelene is a web application that maps moonquakes in interactive 3D. It plots seismic events from the Apollo missions on a model of the Moon.\n\n## The problem\n\nThe Apollo seismometers returned years of data, but it lives in archives formatted for specialists. A map is a faster way to see where the Moon moved.\n\n## What makes it interesting\n\nIt won the Galactic Impact award at NASA Space Apps 2022. The project is an interface over public data, which is a category Space Apps rewards.\n\n## What you can learn\n\nThe same dataset can support several products. A visualization is a legitimate contribution when the raw archive is effectively unusable.\n\n## Limitations of this analysis\n\nThe record does not link the app or describe how the seismic events are geolocated.\n", "description": "Interactive 3D moonquake mapping web application.", "domain": "science", "event": { "award": "Galactic Impact", "name": "NASA Space Apps Challenge", "year": 2022 }, "evidence": [ { "supports": "Space Apps 2022 Galactic Impact award", "url": "https://www.spaceappschallenge.org/blog/2022-global-winners/" } ], "hot": 23, "id": "selene", "name": "Selene", "tags": [ "astronomy", "web-mobile", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "Shaker turned Facebook's social graph into walkable virtual rooms. Disrupt SF 2011 exposure helped it raise serious capital before the virtual-worlds hype cycle cooled.\n", "description": "Virtual bar/lounge environment built on Facebook Open Graph; won TechCrunch Disrupt San Francisco 2011 and raised significant venture funding.", "domain": "social", "event": { "award": "Disrupt Battlefield finalist", "name": "TechCrunch Disrupt San Francisco", "organizer": "TechCrunch", "year": 2011 }, "evidence": [ { "supports": "Shaker was a virtual lounge built on Facebook Open Graph that competed at TechCrunch Disrupt SF 2011 and raised substantial VC", "url": "https://en.wikipedia.org/wiki/Shaker_(social_network)" } ], "hot": 20, "id": "shaker", "links": { "official": "https://en.wikipedia.org/wiki/Shaker_(social_network)" }, "name": "Shaker", "tags": [ "techcrunch-disrupt", "virtual-worlds", "facebook", "social" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nShotgun King: The Final Checkmate is a roguelike chess variant. The player commands the black king, who has a shotgun instead of a full army.\n\n## How the experience works\n\nEach floor presents a white chess formation. The king moves as a king, shoots in a line, and picks up cards that change the rules between runs.\n\n## What makes it interesting\n\nIt was the top-rated entry at Ludum Dare 50 in 2022 and later released on Steam. The premise is one sentence long, which is why it spread.\n\n## What you can learn\n\nA familiar ruleset is free tutorial. Players already know how chess pieces move, so the joke lands without explanation.\n\n## Limitations of this analysis\n\nThe record does not describe the card system or how the balance changed for release.\n", "description": "Chess variant roguelike where one side gets a shotgun. Top-rated entry at Ludum Dare 50.", "domain": "games", "event": { "award": "Top-rated entry", "name": "Ludum Dare 50", "year": 2022 }, "evidence": [ { "supports": "Ludum Dare 50 top-rated placement", "url": "https://ldj.am/games" } ], "hot": 33, "id": "shotgun-king", "links": { "official": "https://store.steampowered.com/app/1968320/Shotgun_King_The_Final_Checkmate/" }, "name": "Shotgun King: The Final Checkmate", "tags": [ "game-jam", "roguelite", "chess" ], "verified_at": "2026-10-24" }, { "body": "A weekend Disrupt hackathon entry that put a flow sensor on a shower and a leaderboard on a phone. Judges awarded it the $5,000 grand prize over Blitz and Interactive Markdown.\n", "description": "Hardware plus mobile app that monitors shower flow rate and gamifies shorter showers against friends; won the TechCrunch Disrupt SF 2014 hackathon grand prize.", "domain": "hardware", "event": { "award": "grand prize", "name": "TechCrunch Disrupt SF Hackathon", "organizer": "TechCrunch", "year": 2014 }, "evidence": [ { "supports": "Hardware hack monitoring shower flow rate with a Go backend and mobile app won Disrupt SF 2014 hackathon grand prize and $5,000", "url": "https://techcrunch.com/2014/09/07/shower-with-friends-wins-the-disrupt-sf-2014-hackathon-grand-prize-blitz-and-interactive-markdown-are-runners-up" } ], "hot": 45, "id": "shower-with-friends", "links": { "demo": "https://techcrunch.com/2014/09/07/shower-with-friends-lets-you-gamify-water-consumption/" }, "name": "Shower With Friends", "tags": [ "hardware", "water", "gamification", "techcrunch-disrupt", "playable" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nSkyShield identifies risk zones for satellites from near-Earth objects and meteor streams. It updates in real time as orbital data changes.\n\n## The problem\n\nSatellite operators need to know when debris or a meteor stream will cross an orbit. The information exists, but it is spread across sources and hard to act on.\n\n## What makes it interesting\n\nIt won Best Use of Science at NASA Space Apps 2024. The product is a risk map, which turns open orbital data into a decision surface.\n\n## What you can learn\n\nFor operational users, a map beats a table. The value is in the aggregation, not the raw data.\n\n## Limitations of this analysis\n\nThe record does not describe the data sources or the update rate.\n", "description": "Identifies satellite risk zones from near-Earth objects and meteor streams in real time.", "domain": "science", "event": { "award": "Best Use of Science", "name": "NASA Space Apps Challenge", "year": 2024 }, "evidence": [ { "supports": "Space Apps 2024 Best Use of Science award", "url": "https://news.sciencex.in/nasa-international-space-apps-challenge-announces-2024-global-winners" } ], "hot": 25, "id": "skyshield", "name": "SkyShield", "tags": [ "space", "safety", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "SkyWatch won Space Apps 2014 by making satellite Earth observation data easier to find and combine, years before similar consumer mapping products.\n", "description": "Space data mashup that won NASA Space Apps 2014 Best Use of Data; combined Earth observation feeds into a single accessible interface.", "domain": "science", "event": { "award": "Best Use of Data", "name": "NASA Space Apps Challenge", "organizer": "NASA", "year": 2014 }, "evidence": [ { "supports": "SkyWatch project page for NASA Space Apps 2014; won Best Use of Data", "url": "https://www.spaceappschallenge.org/2014/challenge/skywatch/" } ], "hot": 20, "id": "skywatch", "links": { "official": "https://www.spaceappschallenge.org/2014/challenge/skywatch/" }, "name": "SkyWatch", "tags": [ "nasa-space-apps", "earth-observation", "data-visualization" ], "verified_at": "2026-10-24" }, { "body": "SolanaFM rode Singapore Management University's startup pipeline from student project to $4.5M seed to a Jupiter Exchange acquisition inside two years.\n", "description": "Solana blockchain explorer and data platform from Singapore Management University's innovation ecosystem; raised US$4.5M seed and was acquired by Jupiter Exchange in 2024.", "domain": "fintech", "event": { "award": "project origin", "name": "SMU Innovation", "organizer": "Singapore Management University", "year": 2023 }, "evidence": [ { "supports": "SMU 2025 Year in Review notes SolanaFM raised US$4.5M seed and was acquired by Jupiter Exchange in 2024; co-founder later recognized in TIME's Best Inventions", "url": "https://iie.smu.edu.sg/sites/iie.smu.edu.sg/files/2025_Year%20in%20Review/2025_Year%20in%20Review_compressed.pdf" } ], "hot": 23, "id": "solana-fm", "links": { "official": "https://solanafm.com" }, "name": "SolanaFM", "tags": [ "solana", "explorer", "singapore", "exit", "student-venture" ], "verified_at": "2026-10-24" }, { "authors": [ "SongSat team" ], "body": "Rather than dump telemetry into another dashboard, SongSat turned it into audio you can actually hear, earning a top global Space Apps 2019 title.\n", "description": "Satellites that transmit data as audible songs for citizen listening; won the 2019 NASA Space Apps global award for Best Scientific Achievement.", "domain": "science", "event": { "award": "Best Scientific Achievement (global winner)", "name": "NASA Space Apps Challenge", "organizer": "NASA", "year": 2019 }, "evidence": [ { "supports": "SongSat listed as a 2019 NASA Space Apps Challenge project transmitting satellite data as audible songs", "url": "https://www.spaceappschallenge.org/2019/challenge/songsat/" }, { "supports": "NASA Space Apps is an international hackathon producing citizen-science tools from space data", "url": "https://www.nasa.gov/space-apps/" } ], "hot": 21, "id": "songsat", "links": { "official": "https://www.spaceappschallenge.org/2019/challenge/songsat/" }, "name": "SongSat", "tags": [ "nasa-space-apps", "space", "sonification", "citizen-science" ], "verified_at": "2026-10-24" }, { "authors": [ "Lunar Outpost" ], "body": "A 2016 Space Apps hardware entry became a commercial lunar-environment sensor family with formal NASA Spinoff status. One of the clearest jam-to-spaceflight-product pipelines.\n", "description": "Radiation and environment sensor network originally developed at NASA Space Apps 2016 as Canaria; evolved into Lunar Outpost's Space Canary product line with NASA Spinoff recognition.", "domain": "science", "event": { "award": "Best Use of Hardware", "name": "NASA Space Apps Challenge", "organizer": "NASA", "year": 2016 }, "evidence": [ { "supports": "NASA Spinoff 2020 profiles Lunar Outpost's Space Canary environmental sensor, noting its Canaria prototype began at a NASA Space Apps event", "url": "https://spinoff.nasa.gov/Spinoff2020/ch_5.html" } ], "hot": 21, "id": "space-canary", "links": { "official": "https://spinoff.nasa.gov/Spinoff2020/ch_5.html" }, "name": "Space Canary (Canaria)", "tags": [ "nasa-space-apps", "space", "hardware", "spinoff" ], "verified_at": "2026-10-24" }, { "body": "## What it does\n\nSphero makes connected rolling robots, including the BB-8 toy. The product is a sealed sphere that drives, lights up, and is controlled from a phone.\n\n## The key insight\n\nThe robot is deliberately dumb. It contains motors, a battery, a radio, and enough firmware to execute commands. The phone holds the interface, the games, and the behavior logic.\n\nThat split means new behavior ships as an app update. The hardware does not need to be redesigned when someone invents a new way to use it, and the bill of materials stays low because the device carries no display and no general-purpose compute.\n\n## How it works\n\nInside the sphere, two wheels drive against the inner shell. Steering comes from the difference in wheel speeds, and the center of mass sits low so the robot self-rights. A sphere with two driven wheels has no front, which is why the robot can move in any direction without turning first.\n\nAn inertial measurement unit reports orientation and acceleration. The firmware uses it to hold a heading and to report collisions. The phone uses it to render motion and to score games.\n\nThe radio link carries short commands: set speed, set color, play a sound, read sensors. The protocol is small because the device is not streaming video or running a language runtime. Latency matters more than bandwidth, since the user is steering.\n\nThe SDK exposes those commands to third-party developers. Sphero's educational line leans on the same idea: the robot is an output device, and the programming happens in an app or a block-based editor.\n\n## Build something similar\n\nBuild a two-wheeled robot in a clear sphere, or simulate one. Expose three commands over a wireless link: drive with a heading and speed, set an LED color, read the IMU.\n\nPut the control loop on the phone. Draw a joystick, map it to a heading and speed, and send the command at a fixed rate.\n\nThen add one behavior that lives entirely in the app, such as a patrol route or a reaction to being picked up. If you can change that behavior without touching firmware, the split is working.\n\n## Sources\n\nStartup Weekend origin and funding are linked in the evidence field.\n", "deep_dive": true, "description": "Connected robot toys including the BB-8 droid. Raised over $120 million.", "domain": "hardware", "event": { "name": "Startup Weekend Boulder", "year": 2010 }, "evidence": [ { "supports": "Startup Weekend Boulder 2010 origin and funding", "url": "https://startupweekend.boston/alumni" } ], "hot": 30, "id": "sphero", "key_idea": "Keep the robot simple and put the intelligence in the phone, so the hardware can stay cheap and the behavior can keep changing.", "learn_from": "Split a hardware product at the boundary that changes fastest. Ship the update to the app, not the device.", "links": { "official": "https://www.sphero.com/" }, "name": "Sphero", "tags": [ "robotics", "consumer", "startup-weekend" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nStarflock tracks how wildfire smoke disperses chemically, using airborne measurement data.\n\n## The problem\n\nSmoke from a large fire travels far and changes composition as it goes. Communities downwind need to know what is arriving, not only where the fire is.\n\n## What makes it interesting\n\nIt won Best Use of Data at NASA Space Apps 2022. The project follows a plume rather than a fire perimeter, which is a different question from most wildfire tools.\n\n## What you can learn\n\nAir quality is a transport problem. The interesting data is often between the source and the sensor.\n\n## Limitations of this analysis\n\nThe record does not describe the measurement sources or the dispersion model.\n", "description": "Tracks chemical dispersion of wildfire smoke using airborne data.", "domain": "climate", "event": { "award": "Best Use of Data", "name": "NASA Space Apps Challenge", "year": 2022 }, "evidence": [ { "supports": "Space Apps 2022 Best Use of Data award", "url": "https://www.spaceappschallenge.org/blog/2022-global-winners/" } ], "hot": 23, "id": "starflock", "name": "Starflock", "tags": [ "wildfire", "data", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nSTARWHAL is a local multiplayer game where narwhals duel with space lances. It supports up to four players on one screen.\n\n## How the experience works\n\nMovement is momentum-based and hard to control. The fun comes from the collision of a clumsy body with a sharp point, which is why matches are short and loud.\n\n## What makes it interesting\n\nIt was created at the Carleton Ottawa site of Global Game Jam 2013, then released commercially and played over a million times across platforms.\n\n## What you can learn\n\nSilly physics is a low-cost way to make a local multiplayer game memorable. The comedy is in the control, not the art.\n\n## Limitations of this analysis\n\nThe record does not describe the porting work or the level design.\n", "description": "Local multiplayer game where narwhals fight with space lances. Over 1 million plays across platforms.", "domain": "games", "event": { "name": "Global Game Jam 2013", "organizer": "Carleton Ottawa site" }, "evidence": [ { "supports": "GGJ 2013 origin and commercial success", "url": "https://globalgamejam.org/ggj-success-stories" } ], "hot": 10, "id": "starwhal", "links": { "official": "https://starwhal.com/" }, "name": "STARWHAL", "tags": [ "game-jam", "local-multiplayer", "indie" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nStorm Prophet is a machine learning model that predicts geomagnetic storm levels. It reads solar wind data from the DSCOVR satellite.\n\n## The problem\n\nA geomagnetic storm can disrupt satellites, GPS, and power grids. Forecasters need lead time, and the solar wind gives only about an hour.\n\n## What makes it interesting\n\nIt won Best Use of Data at NASA Space Apps 2023. The project uses an LSTM on a time series, which is a sensible fit for a signal that unfolds over time.\n\n## What you can learn\n\nIn forecasting, the baseline matters more than the model. A prediction is useful only if it beats the current operational method.\n\n## Limitations of this analysis\n\nThe record does not report accuracy or comparison against existing forecasts.\n", "description": "LSTM model that predicts geomagnetic storm levels from DSCOVR satellite data.", "domain": "science", "event": { "award": "Best Use of Data", "name": "NASA Space Apps Challenge", "year": 2023 }, "evidence": [ { "supports": "Space Apps 2023 Best Use of Data award", "url": "https://www.spaceappschallenge.org/blog/the-nasa-space-apps-challenge-live-global-winners-announcement/" } ], "hot": 23, "id": "storm-prophet", "name": "Storm Prophet", "tags": [ "space-weather", "ai-ml", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "body": "One of the cleanest medical-device pipelines: a weekend health hackathon idea became an FDA-cleared wearable lung monitor used by clinicians and researchers.\n", "description": "RESP Biosensor wearable for real-time lung health monitoring; team formed at a 2016 Independence Blue Cross and Thomas Jefferson University health hackathon, later received FDA 510(k) clearance.", "domain": "healthcare", "event": { "award": "project origin", "name": "Independence Blue Cross Health Hackathon", "organizer": "Independence Blue Cross and Thomas Jefferson University", "year": 2016 }, "evidence": [ { "supports": "Strados team formed at a 2016 Independence Blue Cross and Thomas Jefferson University health hackathon; RESP Biosensor later received FDA 510(k) clearance", "url": "https://technical.ly/startups/wearable-device-lung-health-strados-labs" } ], "hot": 31, "id": "strados-labs", "links": { "official": "https://www.stradoslabs.com" }, "name": "Strados Labs", "tags": [ "wearable", "health-hackathon", "fda", "lung-health" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nThe Street View Trike is a pedicab with a camera mast, built to capture Street View imagery where cars cannot go, such as parks and cultural sites.\n\n## The problem\n\nStreet View covered roads. It could not cover a footpath, a palace courtyard, or a university campus, and those places were a large part of what users searched for.\n\n## What makes it interesting\n\nDan Ratner built it during a Google workshop weekend as a 20% project in 2008. It became a full fleet rather than a one-off demo.\n\n## What you can learn\n\nA data-collection product is limited by its vehicle. Changing the vehicle changes which world is addressable.\n\n## Limitations of this analysis\n\nThe record does not describe the camera rig or the scale of the fleet.\n", "description": "Pedicab camera rig for capturing Street View imagery in parks and cultural sites. Weekend build led to a full fleet.", "domain": "hardware", "event": { "name": "Google Workshops", "organizer": "Google", "year": 2008 }, "evidence": [ { "supports": "Google workshop origin, Dan Ratner, weekend build", "url": "https://googleblog.blogspot.com/2011/04/google-workshops-place-for-googlers-to.html" } ], "hot": 0, "id": "street-view-trike", "name": "Street View Trike", "notes": "Built by Dan Ratner in a 20% project and workshop weekend hack", "tags": [ "maps", "google", "internal-hack-week" ], "verified_at": "2026-10-24" }, { "authors": [ "Brad Menezes" ], "body": "## What it is\n\nSuperblocks is a platform for building internal tools and AI-assisted internal software.\n\n## What makes it interesting\n\nCo-founder Brad Menezes came through Red Bull Basement, a student innovation program, before starting the company in 2021. The record follows that path from a student competition to an enterprise product.\n\n## What you can learn\n\nA student program is a low-cost way to find out whether an idea attracts anyone. It is not proof of a market, but it is a place to learn whether people care.\n\n## Limitations of this analysis\n\nThis record documents the founder's program history, not the product. Superblocks' architecture, pricing, and customer base are not covered by the linked source.\n", "description": "AI-powered internal tools platform; co-founder Brad Menezes came through the Red Bull Basement student innovation program before scaling the company.", "domain": "developer-tools", "event": { "award": "program alumnus", "name": "Red Bull Basement", "organizer": "Red Bull", "year": 2021 }, "evidence": [ { "supports": "Brad Menezes, Red Bull Basement alumnus, co-founded Superblocks in 2021 to help companies build internal tools and AI-powered software", "url": "https://www.instagram.com/reel/DZJbWBgOkpt" } ], "hot": 23, "id": "superblocks", "links": { "official": "https://www.superblocks.com" }, "name": "Superblocks", "tags": [ "red-bull-basement", "internal-tools", "ai", "enterprise" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nSuperfluid is a protocol for streaming money on-chain. Payments flow per second instead of arriving as discrete transfers.\n\n## The problem\n\nSalaries, subscriptions, and vesting are all continuous in concept and discrete in practice. A monthly transfer is an approximation that creates friction and disputes.\n\n## What makes it interesting\n\nIt started at ETHOnline 2020 and became a widely used DeFi primitive. Streaming is a base layer that other applications can build on, which is why it persisted.\n\n## What you can learn\n\nChanging the unit of a payment changes the products built on it. Per-second money makes a different set of applications possible.\n\n## Limitations of this analysis\n\nThe record does not describe the settlement mechanics or the gas model.\n", "description": "Money streaming protocol for continuous payments, subscriptions, and salaries on-chain.", "domain": "fintech", "event": { "name": "ETHOnline", "year": 2020 }, "evidence": [ { "supports": "ETHOnline 2020 origin", "url": "https://bizthon.hashnode.dev/from-hackathon-to-unicorn-real-stories-of-web3-projects-that-made-it-big" } ], "hot": 3, "id": "superfluid", "links": { "official": "https://www.superfluid.finance/" }, "name": "Superfluid", "tags": [ "web3", "defi", "ethereum", "payments" ], "verified_at": "2026-10-24" }, { "authors": [ "SUPERHOT Team" ], "body": "## What it does\n\nSUPERHOT is a first-person shooter where time moves only when the player moves. Stand still and bullets hang in the air. Move and the world resumes. The 2013 jam prototype introduced this mechanic; the full game shipped in 2016.\n\n## The key insight\n\nThe team did not set out to build a shooter. They set out to make something in seven days. The time-movement mechanic emerged from a jam constraint: make one interesting rule and build everything around it.\n\nThe mechanic does double duty. It is a puzzle rule (position yourself before time advances) and a difficulty dial (move less, survive longer). Every level design problem becomes \"what does the player need to do while frozen?\"\n\n## How it works\n\nThe simulation runs on a fixed timestep, but the timestep only advances when the player provides input. The engine has two states: frozen (delta time = 0) and running (delta time = normal).\n\nWhen frozen, physics does not integrate, AI does not tick, and particles pause. Bullets keep their positions because velocity integrates against delta time. Zero delta time freezes them in place.\n\nWhen the player moves (WASD or mouse look), the engine switches to running. One frame of input produces one frame of simulation. When the player stops, the engine drops back to frozen.\n\nThe bullet freezing comes for free. No special-casing required: the mechanic falls out of the physics loop.\n\nLevel design is the real work. Each arena is laid out so the optimal solution involves specific movement sequences: dodge here, shoot there, reposition. The levels teach the mechanic through geometry, not tutorials.\n\n## Build something similar\n\nStart with a simple FPS scene: a player controller, a few enemies that shoot, and bullets that travel.\n\nAdd a pause system. When the player is not pressing movement keys, set time scale to 0. When they press a key, set it back to 1.\n\nTest it: fire a shot, release all keys, watch the bullet stop mid-air.\n\nThen design three arenas around the mechanic. The first should be trivial (stand still, aim, move to shoot). The second should require repositioning between shots. The third should punish standing still (enemies close in over time).\n\n## Sources\n\nPrototype origin and development timeline are documented at superhotgame.com/our-story, linked in the evidence field.\n", "deep_dive": true, "description": "First-person shooter prototype where time moves only when the player moves, built during a seven-day game jam.", "domain": "games", "event": { "name": "7 Day FPS", "year": 2013 }, "evidence": [ { "supports": "Prototype created August 10-17, 2013 during the 7 Day First Person Shooter game jam", "url": "https://superhotgame.com/our-story" }, { "supports": "Original 2013 prototype still playable on the official site", "url": "https://superhotgame.com" } ], "hot": 45, "id": "superhot-prototype", "key_idea": "Time advances only when the player moves, turning a first-person shooter into a puzzle game.", "learn_from": "One mechanical constraint can replace an entire design document. Find the constraint first.", "links": { "demo": "https://superhotgame.com/prototype.html", "official": "https://superhotgame.com" }, "name": "SUPERHOT Prototype", "tags": [ "game-jam", "fps", "unity", "prototype", "playable" ], "verified_at": "2026-10-24" }, { "authors": [ "Bossa Studios" ], "body": "A Global Game Jam joke about clumsy surgery gloves became Bossa Studios' breakout brand, spawning sequels and ports across consoles and VR.\n", "description": "Darkly comic physics-based surgery simulator that started as a Global Game Jam 2013 prototype; became a multi-platform commercial hit.", "domain": "games", "event": { "award": "jam prototype", "name": "Global Game Jam", "organizer": "Global Game Jam", "year": 2013 }, "evidence": [ { "supports": "Surgeon Simulator originated as a Global Game Jam 2013 prototype and expanded into a multi-platform Bossa Studios commercial release", "url": "https://store.steampowered.com/app/233720/Surgeon_Simulator/" } ], "hot": 30, "id": "surgeon-simulator", "links": { "official": "https://store.steampowered.com/app/233720/Surgeon_Simulator/" }, "name": "Surgeon Simulator", "tags": [ "game-jam", "physics", "comedy", "steam", "ggj" ], "verified_at": "2026-10-24" }, { "body": "Swirl W@tch is a straightforward jam-to-Steam case: a 48-hour 2021 prototype that got a multi-year polish pass before commercial release.\n", "description": "Game made in 48 hours at Global Game Jam 2021; the full polished version shipped on Steam in April 2023 with a positive review rating.", "domain": "games", "event": { "name": "Global Game Jam 2021", "organizer": "Global Game Jam", "year": 2021 }, "evidence": [ { "supports": "GGJ news: original Swirl W@tch made in 48 hours at Global Game Jam 2021; full polished game published on Steam April 2023 with positive reviews", "url": "https://globalgamejam.org/news/jam-fame-global-game-jam-projects-became-successful-published-games" } ], "hot": 13, "id": "swirl-watch", "name": "Swirl W@tch", "tags": [ "global-game-jam", "commercial", "indie" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nTakeOne is a robotic film studio. It plans shots, moves the camera and lighting, and edits the footage.\n\n## The problem\n\nA small production has a crew problem. A single operator can hold a camera or a light, not both, and coverage is what separates a professional look from a home video.\n\n## What makes it interesting\n\nIt was a finalist at Hack the North 2026. The project treats the whole shoot as one system, from planning to edit, rather than automating a single camera move.\n\n## What you can learn\n\nRobotics in creative work is a workflow problem. The robot has to fit the way a shoot already runs.\n\n## Limitations of this analysis\n\nThe record does not describe the hardware or the planning model.\n", "description": "AI-powered robotic film studio that plans shots, moves camera and lighting, and edits footage.", "domain": "media-arts", "event": { "name": "Hack the North", "year": 2026 }, "evidence": [ { "supports": "Hack the North 2026 finalist", "url": "https://museum.hackthenorth.com/" } ], "hot": 5, "id": "takeone", "name": "TakeOne", "tags": [ "robotics", "ai-ml", "film", "college-hackathon" ], "verified_at": "2026-10-24" }, { "body": "Tiago Paiva built the first version at a Twilio hackathon in 2011 after watching his mother struggle with traditional call-center software. It grew into a public company.\n", "description": "Cloud-based contact center platform that lets businesses handle customer calls over the internet instead of traditional phone systems.", "domain": "web-mobile", "event": { "name": "TwilioCon hackathon", "year": 2011 }, "evidence": [ { "supports": "Talkdesk began as a hackathon project", "url": "https://www.talkdesk.com/about" } ], "hot": 10, "id": "talkdesk", "links": { "official": "https://www.talkdesk.com" }, "name": "Talkdesk", "tags": [ "saas", "customer-service", "cloud" ], "verified_at": "2026-10-24" }, { "body": "Tanzaku shared the top Art Bytes 2012 prize with Frame, showing how museum hackathons can yield multiple polished mobile experiences from one weekend.\n", "description": "Smartphone museum exploration app; co-winner of the Walters Art Museum's first Art Bytes hackathon in 2012 with a $1,000 prize.", "domain": "media-arts", "event": { "award": "co-winner ($1,000)", "name": "Art Bytes", "organizer": "Walters Art Museum", "year": 2012 }, "evidence": [ { "supports": "Tanzaku and Frame each took top $1,000 prizes at the first Walters Art Museum Art Bytes hackathon in 2012", "url": "https://technical.ly/software-development/art-bytes-these-are-the-eight-projects-from-the-first-ever-walters-art-museum-hackathon" } ], "hot": 20, "id": "tanzaku", "links": { "official": "https://technical.ly/baltimore/2012/07/27/art-bytes-brings-technologists-others-to-walters-art-museum-for-art-inspired-hackathon/" }, "name": "Tanzaku", "tags": [ "museum-hackathon", "mobile", "art", "walters" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nTerra Nil is a reverse city builder. The player restores a wasteland into a working ecosystem, then removes their own machinery and leaves.\n\n## How the experience works\n\nBuildings purify soil and water and reintroduce species. Once the biome sustains itself, the player recycles the equipment, which is the opposite of a normal builder's ending.\n\n## What makes it interesting\n\nIt began as a Ludum Dare 2019 prototype and later released commercially, including on Netflix Games.\n\n## What you can learn\n\nInverting a genre's goal changes the whole emotional register. The win condition is the absence of the player's work.\n\n## Limitations of this analysis\n\nThe record does not describe the ecology model or how the jam version changed.\n", "description": "Reverse city builder about ecosystem restoration. Released on Netflix Games.", "domain": "games", "event": { "name": "Ludum Dare", "year": 2019 }, "evidence": [ { "supports": "Ludum Dare prototype origin", "url": "https://itch.io/games/top-sellers/tag-ludum-dare" } ], "hot": 11, "id": "terra-nil", "links": { "official": "https://www.terraformgame.com/" }, "name": "Terra Nil", "tags": [ "game-jam", "strategy", "indie" ], "verified_at": "2026-10-24" }, { "authors": [ "Matteo Nitto", "Dario Fazio" ], "body": "Most jam games stay jam games. The Textorcist's typing-bullet-hell hybrid survived the jump to a polished Steam and Nintendo Switch release, a rare jam-to-shelf path.\n", "description": "Typing-action game where you fight demons by typing scripture; born from a Global Game Jam 2016 prototype, later shipped on Steam and Switch.", "domain": "games", "event": { "award": "jam prototype", "name": "Global Game Jam", "organizer": "Global Game Jam", "year": 2016 }, "evidence": [ { "supports": "The Textorcist began as a Global Game Jam 2016 prototype and was expanded into a full commercial release", "url": "https://www.gamedeveloper.com/design/making-the-textorcist-how-one-jam-game-became-a-full-steam-release" } ], "hot": 31, "id": "textorcist", "links": { "official": "https://store.steampowered.com/app/774361/The_Textorcist/" }, "name": "The Textorcist", "tags": [ "game-jam", "typing", "action", "ggj" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nThe Box is decentralized physical infrastructure for peer-to-peer package delivery. It combines custom hardware with on-chain coordination.\n\n## The problem\n\nA package moves through a courier network that the sender does not control. A peer-to-peer option could use capacity that already exists, such as a neighbor already driving that route.\n\n## What makes it interesting\n\nIt won the Hardware and IoT DePIN category at ETHGlobal Brussels 2024. The project is a physical device plus a protocol, which is a much harder demo than software alone.\n\n## What you can learn\n\nDePIN projects live or die on the hardware cost. A box that has to be manufactured and shipped changes the economics of the network.\n\n## Limitations of this analysis\n\nThe record does not describe the lock mechanism, the chain integration, or any pilot.\n", "description": "Decentralized physical infrastructure for peer-to-peer package delivery with custom hardware.", "domain": "hardware", "event": { "award": "Hardware IoT DePIN winner", "name": "ETHGlobal Brussels", "year": 2024 }, "evidence": [ { "supports": "ETHGlobal Brussels 2024 Hardware IoT DePIN winner", "url": "https://www.recursive.so/hackathons" } ], "hot": 25, "id": "the-box", "name": "The Box", "tags": [ "web3", "hardware", "depin", "logistics" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nThe King is Watching is a roguelite where production only happens while the king is looking. Crops grow and miners work under his gaze, and stop when he looks away.\n\n## How the experience works\n\nThe player decides where to direct attention. Since one gaze cannot cover the whole kingdom, every run is a choice about what to neglect.\n\n## What makes it interesting\n\nIt was a top-rated entry at Ludum Dare 54 and later sold over 600,000 copies on Steam with tinybuild publishing.\n\n## What you can learn\n\nA single restriction can be the whole design. \"Only while watched\" generates every decision without extra systems.\n\n## Limitations of this analysis\n\nThe record does not describe how the attention meter is tuned or what the full release added.\n", "description": "Roguelite where crops and miners produce only while the king watches. Sold 600k+ copies on Steam.", "domain": "games", "event": { "award": "Top-rated entry", "name": "Ludum Dare 54" }, "evidence": [ { "supports": "Ludum Dare 54 origin, tinybuild publishing, sales figures", "url": "https://gamemaker.io/en/blog/the-king-is-watching-tinybuild" } ], "hot": 30, "id": "the-king-is-watching", "links": { "official": "https://store.steampowered.com/app/2401630/The_King_is_Watching/" }, "name": "The King is Watching", "tags": [ "game-jam", "roguelite", "indie" ], "verified_at": "2026-10-24" }, { "body": "Daniel Linssen (Managore) won Ludum Dare 29 Overall with this prototype. The full version later shipped on Steam and other platforms.\n", "description": "Puzzle platformer with a unique mechanic letting the player dive into and travel through the ground.", "domain": "games", "event": { "name": "Ludum Dare 29", "year": 2014 }, "evidence": [ { "supports": "Developer Daniel Linssen's own page: made for Ludum Dare #29", "url": "https://managore.wordpress.com/games/the-sun-and-moon/" }, { "supports": "Development log for Ludum Dare #29 where The Sun And Moon was created", "url": "https://managore.wordpress.com/2014/05/15/ludum-dare-29/" } ], "hot": 10, "id": "the-sun-and-moon", "links": { "official": "https://managore.itch.io/the-sun-and-moon" }, "name": "The Sun and Moon", "tags": [ "puzzle", "platformer", "precision" ], "verified_at": "2026-10-24" }, { "body": "Acid Nerve (Mark Foster and David Fenn) built the Flash prototype in a weekend. The full game expanded to sixteen titans and launched April 2015 via Devolver Digital.\n", "description": "Minimalist boss-rush action game where the player has one arrow and one life, inspired by Shadow of the Colossus and Dark Souls.", "domain": "games", "event": { "name": "Ludum Dare 28", "year": 2013 }, "evidence": [ { "supports": "Originally created for Ludum Dare 28 as TITAN SOULS; placed first in Overall, Audio, and Mood (Jam) categories", "url": "https://en.wikipedia.org/wiki/Titan_Souls" }, { "supports": "Original jam prototype page by Acid Nerve, made for Ludum Dare 28 theme 'You Only Get One'", "url": "https://clawmark.itch.io/titan-souls" }, { "supports": "Acid Nerve created Titan Souls during the weekend-long Ludum Dare game jam #28; shown at E3 six months later, published by Devolver Digital", "url": "https://www.escapistmagazine.com/deaths-door-interview-acid-nerve-mark-foster-david-fenn/" } ], "hot": 10, "id": "titan-souls", "links": { "official": "https://store.steampowered.com/app/297130/Titan_Souls/" }, "name": "Titan Souls", "tags": [ "action", "boss-rush", "pixel-art" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nTowerFall is a local multiplayer archery game. Players fire arrows at each other in single-screen arenas, where one hit ends a round and arrows can be picked up after they land.\n\n## The problem\n\nLocal multiplayer in 2012 was mostly party games built for casual play. TowerFall aimed at something faster and more precise, with the tension of a fighting game and the readability of a platformer.\n\n## What makes it interesting\n\nIt started at the Vancouver Full Indie Game Jam in 2012. It became the first commercial release from Maddy Thorson, who later made Celeste.\n\n## What you can learn\n\nA jam build can be the whole design. TowerFall's rules fit on an index card: one hit, limited arrows, pick them up. The depth came from arena layout and movement, not from added systems.\n\n## Limitations of this analysis\n\nThe record documents the jam origin and the designer, not the netcode or the changes between the jam build and the commercial release.\n", "description": "Multiplayer archer combat game. First commercial release of Maddy Thorson, who later made Celeste.", "domain": "games", "event": { "name": "Vancouver Full Indie Game Jam", "year": 2012 }, "evidence": [ { "supports": "Game jam origin and Maddy Thorson connection", "url": "https://gamemaker.io/en/blog/best-game-jam-games" } ], "hot": 10, "id": "towerfall", "links": { "official": "https://towerfall-game.com/" }, "name": "TowerFall", "tags": [ "game-jam", "local-multiplayer", "indie" ], "verified_at": "2026-10-24" }, { "body": "Portuguese studio Whales And Games built the jam version during the pandemic-era digital GGJ 2021, then expanded it into a full commercial release.\n", "description": "Relaxing exploration-trading game where players fly an airship across a world discovering towns and trading goods.", "domain": "games", "event": { "name": "Global Game Jam 2021", "year": 2021 }, "evidence": [ { "supports": "Townseek originally created for Global Game Jam 2021; listed as a GGJ success story", "url": "https://globalgamejam.org/ggj-success-stories" }, { "supports": "Whales And Games: 'Townseek was first created for Global Game Jam 2021'", "url": "https://whalesandgames.com/games/townseek" } ], "hot": 13, "id": "townseek", "links": { "official": "https://store.steampowered.com/app/1659360/Townseek/" }, "name": "Townseek", "tags": [ "exploration", "trading", "relaxing" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nTrello organizes work into boards, lists, and cards. A card is a task; moving it between lists shows progress. The board is a kanban wall that a team can read without training.\n\n## The problem\n\nProject tools of the time were built around fields, filters, and reports. A small team that wanted to see who was doing what had to configure a system before using it. Trello's bet was that the board metaphor already existed in offices, on walls and sticky notes.\n\n## What makes it interesting\n\nIt came out of Fog Creek's Creek Weeks, an internal hack-week format where employees pitched and prototyped products. The prototype was pitched in January 2011 and the product launched at TechCrunch Disrupt that September. Closing that gap is the part most hack-week projects never manage.\n\n## What you can learn\n\nA familiar physical object is a cheaper interface than a new vocabulary. If users can guess where to click, onboarding is mostly done.\n\n## Limitations of this analysis\n\nThe record documents the origin and launch, not the implementation. Trello's data model, sync approach, and early scaling work are not covered by the linked sources.\n", "description": "Visual collaboration tool that organizes projects into boards, lists, and cards.", "domain": "web-mobile", "event": { "name": "Fog Creek Creek Week", "organizer": "Fog Creek Software", "year": 2011 }, "evidence": [ { "supports": "Originated in Fog Creek's internal Creek Weeks product explorations; prototype pitched January 2011", "url": "https://www.trello.com/about" }, { "supports": "Fog Creek's internal product initiative that became Trello", "url": "https://www.joelonsoftware.com/2011/09/13/announcing-trello" } ], "hot": 0, "id": "trello", "links": { "official": "https://trello.com" }, "name": "Trello", "tags": [ "project-management", "collaboration", "kanban" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nTripLingo is a translation app for travelers. It combines phrasebooks, slang, and a live translator.\n\n## The problem\n\nA phrasebook gives you correct sentences you will not hear back. A traveler needs to understand the reply, including the informal speech a textbook omits.\n\n## What makes it interesting\n\nIt started at Startup Weekend Atlanta and grew into a company. The product treated slang and culture as part of the language problem.\n\n## What you can learn\n\nTranslation is two problems, speaking and understanding. Most early apps solved only the first.\n\n## Limitations of this analysis\n\nThe record does not describe the translation stack or the company's later status.\n", "description": "Traveler translation app with phrasebooks, slang, and live translator features.", "domain": "web-mobile", "event": { "name": "Startup Weekend Atlanta" }, "evidence": [ { "supports": "Startup Weekend Atlanta origin", "url": "https://www.inc.com/articles/201111/companies-that-launched-in-three-days-startup-weekend.html" } ], "hot": 10, "id": "triplingo", "links": { "official": "https://www.triplingo.com/" }, "name": "TripLingo", "tags": [ "travel", "translation", "startup-weekend" ], "verified_at": "2026-10-24" }, { "authors": [ "Jack Dorsey", "Florian Weber" ], "body": "## What it is\n\nTwitter is a public microblogging service. A post is short, visible by default, and read in a timeline.\n\n## The problem\n\nBlogging required time and a title. Status updates existed inside social networks, but not as a public, followable stream.\n\n## How the experience works\n\nA user posts a short message, and followers see it in order. The service began as an SMS product, so a post had to fit a text message.\n\n## What makes it interesting\n\nIt came out of a daylong brainstorming session at Odeo in 2006, and Jack Dorsey and Florian Weber built the first prototype. The first tweet went out on March 21, 2006. The length limit was a carrier constraint that became the product's identity.\n\n## What you can learn\n\nA constraint imposed by infrastructure can become the feature users ask for. Removing it later is a product decision, not a cleanup.\n\n## Limitations of this analysis\n\nThe record covers the origin and the first tweet, not the architecture or later changes to the length limit.\n", "description": "Microblogging service for short public status updates, prototyped inside the podcasting company Odeo.", "domain": "web-mobile", "event": { "name": "Odeo internal brainstorming session", "organizer": "Odeo", "year": 2006 }, "evidence": [ { "supports": "Coding began March 13, 2006; first tweet March 21, 2006", "url": "https://www.wired.com/2011/03/0321twitter-first-tweet" }, { "supports": "Started as a hobby project inside Odeo; Jack Dorsey wrote the first code", "url": "https://www.latimes.com/archives/blogs/technology-blog/story/2011-03-13/twitter-came-to-life-five-years-ago-this-week-creator-jack-dorsey-remembers" }, { "supports": "Origins in a daylong brainstorming session at Odeo; first prototype by Dorsey and Weber", "url": "https://en.wikipedia.org/wiki/History_of_Twitter" } ], "hot": 0, "id": "twitter", "links": { "official": "https://twitter.com" }, "name": "Twitter", "tags": [ "social", "sms", "status-updates" ], "verified_at": "2026-10-24" }, { "body": "Persistence pays: after four hackathon attempts, Unrugable's Solana hardware wallet took Colosseum's grand prize, which includes a direct venture-fund investment pathway.\n", "description": "Solana-native hardware wallet with companion app; crowned grand champion of the Colosseum Cypherpunk Hackathon at Breakpoint 2025 after four prior hackathon attempts.", "domain": "fintech", "event": { "award": "grand champion", "name": "Colosseum Cypherpunk Hackathon", "organizer": "Colosseum", "year": 2025 }, "evidence": [ { "supports": "Unrugable crowned grand champion of Colosseum's Cypherpunk Hackathon at Breakpoint 2025; top 10 winners receive Colosseum venture fund investment and accelerator participation", "url": "https://solanacompass.com/learn/breakpoint-25/winners-of-the-colosseum-cypherpunk-hackathon" } ], "hot": 35, "id": "unrugable", "links": { "official": "https://solanacompass.com/learn/breakpoint-25/winners-of-the-colosseum-cypherpunk-hackathon" }, "name": "Unrugable", "tags": [ "solana", "hardware-wallet", "colosseum", "hackathon-winner" ], "verified_at": "2026-10-24" }, { "authors": [ "Ory Okolloh", "Erik Hersman", "Juliana Rotich", "David Kobia" ], "body": "## What it is\n\nUshahidi is an open-source platform for collecting eyewitness reports and placing them on a map. Reports arrive by SMS or the web and appear as points that responders can filter.\n\n## The problem\n\nDuring the 2008 post-election violence in Kenya, information was spread across blogs, news, and phone calls. There was no shared map of what was happening where.\n\n## How the experience works\n\nA report is sent by text message or submitted online. The platform geocodes it, adds it to the map, and lets moderators review and categorize it. The map is the interface; the reports are the data.\n\n## What makes it interesting\n\nFour technologists built it in days as an ad hoc response, not a planned product. It won the NetSquared Mashup Challenge in 2008 and became a platform used in later crises. The SMS input was the key choice: the people closest to an event often have the weakest connection.\n\n## What you can learn\n\nWhen the network is unreliable, the input channel decides whether the tool works at all. Design for the phone people have, not the one you wish they had.\n\n## Limitations of this analysis\n\nThe record does not cover the moderation workflow, verification methods, or later versions of the platform.\n", "description": "Open-source crisis-mapping platform that crowdsources and maps eyewitness reports via SMS and the web.", "domain": "civic-tech", "event": { "award": "NetSquared Mashup Challenge winner", "name": "Kenya post-election crisis rapid build", "year": 2008 }, "evidence": [ { "supports": "Born out of the 2008 Kenyan post-election violence as an ad hoc rapid software build", "url": "https://www.theguardian.com/technology/2010/aug/11/ushahidi-crowdmap-crisis-mapping" }, { "supports": "Initial deployment January 2008; won the NetSquared mashup challenge May 2008", "url": "https://www.ushahidi.com/about/blog/ushahidi-1-year-later" }, { "supports": "Founded in 2008; built in days by four technologists", "url": "https://en.wikipedia.org/wiki/Ushahidi" } ], "hot": 35, "id": "ushahidi", "links": { "official": "https://www.ushahidi.com", "source": "https://github.com/ushahidi/platform" }, "name": "Ushahidi", "tags": [ "crisis-mapping", "crowdsourcing", "open-source", "maps" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nThe Voyager Transponder is a 200-gigabit optical transponder prototype. Network engineers built it at a Facebook hackathon in 2016.\n\n## The problem\n\nLong-haul optical links are expensive because the equipment is proprietary and specialized. A cheaper, open design could lower the cost of backbone capacity.\n\n## What makes it interesting\n\nThe prototype contributed to the Telecom Infra Project, an effort to open up telecom hardware. A hackathon build feeding a industry consortium is an unusual path.\n\n## What you can learn\n\nInternal hackathons can seed external standards work. The prototype becomes an argument for a different supply chain.\n\n## Limitations of this analysis\n\nThe record does not describe the prototype's performance or what happened to the design.\n", "description": "200G optical transponder prototype built by network engineers at a hackathon. Contributed to the Telecom Infra Project.", "domain": "hardware", "event": { "name": "Facebook Hackathon", "organizer": "Facebook", "year": 2016 }, "evidence": [ { "supports": "Facebook 2016 hackathon, Telecom Infra Project contribution", "url": "https://engineering.fb.com/2016/12/29/developer-tools/favorite-hacks-of-2016/" } ], "hot": 11, "id": "voyager-transponder", "name": "Voyager Transponder", "tags": [ "networking", "optical", "facebook", "internal-hack-week" ], "verified_at": "2026-10-24" }, { "authors": [ "Ricky Wong" ], "body": "Vurb's contextual mobile search won Disrupt's Battlestage and later fetched roughly $300M from Rakuten, one of the larger Disrupt-to-exit outcomes.\n", "description": "Contextual search and discovery app that won the TechCrunch Disrupt 2014 Battlestage; later acquired by Rakuten for approximately $300 million.", "domain": "web-mobile", "event": { "award": "Battlestage winner", "name": "TechCrunch Disrupt", "organizer": "TechCrunch", "year": 2014 }, "evidence": [ { "supports": "Vurb won TechCrunch Disrupt 2014 Battlestage and was acquired by Rakuten for approximately $300 million", "url": "https://en.wikipedia.org/wiki/Vurb" } ], "hot": 20, "id": "vurb", "links": { "official": "https://en.wikipedia.org/wiki/Vurb" }, "name": "Vurb", "tags": [ "techcrunch-disrupt", "search", "mobile", "exit" ], "verified_at": "2026-10-24" }, { "body": "Days after GPT-5's release, Waddle took the top slot at the OpenAI/Cerebral Valley weekend hackathon with an ecommerce digital clone system, then leveraged the win into enterprise traction.\n", "description": "AI-powered digital clone and sales strategy simulation system for ecommerce; won first place among 93 global teams at the OpenAI GPT-5 Hackathon 2025.", "domain": "ai-ml", "event": { "award": "first place", "name": "OpenAI GPT-5 Hackathon", "organizer": "OpenAI and Cerebral Valley", "year": 2025 }, "evidence": [ { "supports": "Waddle won first place among 93 global teams at the OpenAI GPT-5 Hackathon in San Francisco, August 9-10 2025, with a GPT-5-powered digital clone and sales strategy simulation for ecommerce", "url": "https://blog.gentooai.com/waddle-wins-first-place-at-the-openai-gpt-5-hackathon/" } ], "hot": 35, "id": "waddle", "links": { "official": "https://waddle.ai" }, "name": "Waddle", "tags": [ "openai", "ai-agents", "ecommerce", "hackathon-winner" ], "verified_at": "2026-10-24" }, { "body": "Vlambeer (Jan Willem Nijman and Rami Ismail) built the prototype in the Mojam charity jam with Notch and Humble Bundle. It was rebranded Nuclear Throne and shipped in 2014.\n", "description": "Top-down roguelike shooter where mutants fight through procedurally generated levels to claim the Nuclear Throne.", "domain": "games", "event": { "name": "Mojam 2013", "year": 2013 }, "evidence": [ { "supports": "Began as a game jam prototype called Wasteland Kings in February 2013; created during the 48-hour Mojam charity game jam run by Mojang and Humble Bundle", "url": "https://en.wikipedia.org/wiki/Nuclear_Throne" }, { "supports": "Original prototype page: 'the prototype that Vlambeer created for the three day MOJAM charity build'", "url": "https://vlambeer.itch.io/wasteland-kings" }, { "supports": "Rami Ismail interview confirming the Mojam 2013 origin of Wasteland Kings", "url": "https://www.polygon.com/2013/8/15/4625784/vlambeers-wasteland-kings-rami-ismail-interview" } ], "hot": 10, "id": "wasteland-kings", "links": { "official": "https://vlambeer.itch.io/wasteland-kings" }, "name": "Wasteland Kings (Nuclear Throne)", "tags": [ "roguelike", "shooter", "procedural" ], "verified_at": "2026-10-24" }, { "body": "A weekend anti-cheat prototype became a real classroom signal, flagging students who left the quiz tab during assessments at Wayground scale.\n", "description": "Browser-based quiz anti-cheat built during a Wayground internal hackathon; deployed across millions of quiz sessions to deter tab-switching and second-screen cheating.", "domain": "education", "event": { "name": "Wayground Internal Hackathon", "organizer": "Wayground" }, "evidence": [ { "supports": "Wayground internal hackathon produced the Eye See You anti-cheat feature later used across millions of quiz sessions", "url": "https://wayground.com/blog/hackathon-anti-cheat" } ], "hot": 10, "id": "wayground-eye-see-you", "links": { "official": "https://www.wayground.com" }, "name": "Wayground Anti-Cheat \"Eye See You\"", "tags": [ "internal-hackathon", "edtech", "anti-cheat" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nWhat's New? is a machine learning pipeline that estimates the probability of a Carrington-class solar storm.\n\n## The problem\n\nA storm of that size would damage satellites and grids. It is rare, so any estimate has to handle a heavily imbalanced dataset.\n\n## What makes it interesting\n\nIt won Best Use of Science at NASA Space Apps 2022. The project frames the task as a probability rather than a yes-or-no forecast, which suits a rare event.\n\n## What you can learn\n\nRare-event prediction needs a calibration story. A model that says \"probably not\" every day is right most days and useless.\n\n## Limitations of this analysis\n\nThe record does not describe the features, the training data, or the model's calibration.\n", "description": "Machine learning pipeline predicting the probability of a Carrington-class solar storm.", "domain": "science", "event": { "award": "Best Use of Science", "name": "NASA Space Apps Challenge", "year": 2022 }, "evidence": [ { "supports": "Space Apps 2022 Best Use of Science award", "url": "https://www.spaceappschallenge.org/blog/2022-global-winners/" } ], "hot": 23, "id": "whats-new-solar-storm", "name": "What's New?", "tags": [ "space-weather", "ai-ml", "nasa-space-apps" ], "verified_at": "2026-10-24" }, { "authors": [ "Marinos" ], "body": "Disrupt NY 2015's hackathon prize went to Witness, a Twilio-powered crisis reporting app assembled over the event weekend.\n", "description": "Crisis-reporting mobile app built to win the TechCrunch Disrupt New York 2015 Hackathon; used Twilio and crowd-sourced field reports for real-time safety data.", "domain": "civic-tech", "event": { "award": "winner", "name": "TechCrunch Disrupt New York Hackathon", "organizer": "TechCrunch", "year": 2015 }, "evidence": [ { "supports": "Witness wins the TechCrunch Disrupt NY 2015 Hackathon with a Twilio-powered crisis reporting app", "url": "https://techcrunch.com/2015/05/06/witness-wins-the-techcrunch-disrupt-ny-2015-hackathon/" } ], "hot": 31, "id": "witness-disrupt", "links": { "official": "https://techcrunch.com/2015/05/06/witness-wins-the-techcrunch-disrupt-ny-2015-hackathon/" }, "name": "Witness", "tags": [ "techcrunch-disrupt", "crisis", "twilio", "mobile" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nWorkflow is a visual automation app for iOS. Users chain actions into a script and run it with one tap.\n\n## The problem\n\nPhones can do many things, but each task requires a sequence of app switches. Automation lets a user define the sequence once.\n\n## What makes it interesting\n\nAri Weinstein started it after MHacks 2014 at the University of Michigan. Apple acquired it in 2016 and turned it into Shortcuts.\n\n## What you can learn\n\nA hackathon project can become a platform feature. The acquisition was a shortcut to distribution no independent app could match.\n\n## Limitations of this analysis\n\nThe record does not describe the app's action model or the terms of the acquisition.\n", "description": "Visual automation app for iOS. Acquired by Apple in 2016 and became Shortcuts.", "domain": "web-mobile", "event": { "name": "MHacks", "organizer": "University of Michigan", "year": 2014 }, "evidence": [ { "supports": "MHacks 2014 origin, Ari Weinstein, Apple acquisition", "url": "https://medium.com/rough-draft-ventures/from-rough-draft-to-forbes-a-flashback-to-forbes-under-30-honorees-first-photos-products-e1a40589bce1" } ], "hot": 0, "id": "workflow-shortcuts", "links": { "official": "https://apps.apple.com/app/shortcuts/id913512399" }, "name": "Workflow", "tags": [ "ios", "automation", "college-hackathon" ], "verified_at": "2026-10-24" }, { "body": "Yesterday's News appears on Global Game Jam's official success-stories list as a jam-to-Steam path. Exact jam year was not pinned in the sources reviewed.\n", "description": "Game that started as a Global Game Jam entry and was later developed into a Steam release.", "domain": "games", "event": { "name": "Global Game Jam", "organizer": "Global Game Jam" }, "evidence": [ { "supports": "GGJ success-stories page lists Yesterday's News as starting as a jam project before a Steam release", "url": "https://globalgamejam.org/ggj-success-stories" } ], "hot": 10, "id": "yesterdays-news", "links": { "official": "https://store.steampowered.com/app/2850560/Yesterdays_News/" }, "name": "Yesterday's News", "tags": [ "global-game-jam", "commercial", "indie" ], "verified_at": "2026-10-24" }, { "authors": [ "Art Agrawal", "Dongyi Liao" ], "body": "Art Agrawal and Dongyi Liao won TechCrunch Disrupt's Startup Battlefield in 2012 with an online auto repair referral service. The company branded as YourMechanic and scaled a mobile-mechanic marketplace.\n", "description": "Marketplace connecting car owners with mobile mechanics; won the TechCrunch Disrupt SF 2012 Startup Battlefield after pitching earlier as HiretoFix.", "domain": "other", "event": { "award": "winner", "name": "TechCrunch Disrupt SF Startup Battlefield", "organizer": "TechCrunch", "year": 2012 }, "evidence": [ { "supports": "YourMechanic took the TechCrunch Disrupt Cup in September 2012; founders previously operated as HiretoFix", "url": "https://www.bizjournals.com/sanjose/blog/2012/09/yourmechanic-takes-techcrunch-disrupt.html" } ], "hot": 20, "id": "your-mechanic", "links": { "official": "https://www.yourmechanic.com" }, "name": "YourMechanic", "tags": [ "marketplace", "automotive", "techcrunch-disrupt" ], "verified_at": "2026-10-24" }, { "authors": [ "Bo Fishback", "Eric Koester", "Ian Hunter" ], "body": "A last-minute pitch at Startup Weekend LA became a national talking-point startup in under a month, drawing celebrity investors and a board seat for a former eBay CEO.\n", "description": "Hyperlocal real-time buyer-powered marketplace; prototyped and pitched at Startup Weekend Los Angeles in 2011, then raised a $14M Series A led by Kleiner Perkins.", "domain": "web-mobile", "event": { "award": "winner", "name": "Startup Weekend Los Angeles", "organizer": "Startup Weekend", "year": 2011 }, "evidence": [ { "supports": "Within six hours of the 54-hour LA Startup Weekend in February 2011, the founders cooked up Zaarly and built the first prototype", "url": "https://www.entrepreneur.com/leadership/from-business-idea-to-successful-startup-in-three-months/220019" }, { "supports": "Zaarly raised $14.1M Series A from Kleiner Perkins and Sands Capital; Ashton Kutcher seed investor; Meg Whitman joined board", "url": "https://www.businessinsider.com/zaarly-raises-14-million-2011-10" }, { "supports": "Zaarly was little more than an idea when it won LA Startup Weekend three weeks prior to launch coverage", "url": "https://techcrunch.com/2011/03/09/want-someone-to-bring-you-a-beer-get-that-and-anything-else-from-zaarly" } ], "hot": 30, "id": "zaarly", "links": { "official": "https://en.wikipedia.org/wiki/Zaarly" }, "name": "Zaarly", "tags": [ "marketplace", "startup-weekend", "sharing-economy" ], "verified_at": "2026-10-24" }, { "authors": [ "Wade Foster", "Bryan Helmig", "Mike Knoop" ], "body": "## What it does\n\nZapier connects over 7,000 web applications. A Zap is a trigger (new email, new row, new file) paired with an action (create a task, send a message, add a row) in a different app. When the trigger fires, Zapier performs the action.\n\n## The key insight\n\nZapier started at a Startup Weekend in Columbia, Missouri in October 2011. Wade Foster, Mike Knoop, and Bryan Vance built a demo that connected two apps in a weekend. The insight that turned it into a company: every app has an API, but most app pairs do not have a direct integration. A middleware layer that translates between thousands of APIs is more valuable than building any single integration.\n\nThe product is not any one Zap. It is the platform that lets a non-developer create the next thousand without writing code.\n\n## How it works\n\n**Trigger.** Zapier polls an app's API (or receives a webhook) for new events. When a new Gmail arrives, a new row appears in Sheets, or a new file lands in Dropbox, the trigger fires.\n\n**Mapping.** The user defines which fields from the trigger map to which fields in the action. Gmail Subject becomes Slack Message Text. This is a simple key-value transformation.\n\n**Action.** Zapier calls the target app's API with the mapped data. If the API returns an error, Zapier retries with exponential backoff.\n\n**Polling vs webhooks.** Some apps support webhooks (push). Others only support polling (pull). Zapier handles both; the user does not know which is happening.\n\nThe engineering challenge is coverage. Each of the 7,000 apps has its own auth scheme, rate limits, pagination style, and error format. Zapier maintains an adapter for each. The platform value grows with every adapter added.\n\n## Build something similar\n\nBuild a mini-Zapier that connects two apps you use daily.\n\nPick a trigger app with a simple API (GitHub: new commit, or RSS feed: new item). Pick an action app with a simple API (Slack: post message, or Notion: create page).\n\nWrite a script that polls the trigger every 60 seconds, checks for new items since the last poll, and calls the action API for each new item. Store the last-seen timestamp in a file.\n\nThis is one Zap. The platform insight comes when you abstract the trigger and action behind interfaces and let users define the mapping. The second and third Zaps take the same code as the first.\n\n## Sources\n\nStartup Weekend origin is documented in Silicon Prairie News and Inc. articles linked in the evidence field.\n", "deep_dive": true, "description": "Workflow automation platform that connects web apps so users can move data between services without writing code.", "domain": "developer-tools", "event": { "award": "Winner", "name": "Startup Weekend Columbia", "organizer": "Startup Weekend", "year": 2011 }, "evidence": [ { "supports": "Winner of the first Startup Weekend in Columbia, Missouri (October 2011)", "url": "https://zapier.com/blog/zapier-100-connected-services-year-review/" }, { "supports": "Co-founded at the Missouri Startup Weekend", "url": "https://www.ycombinator.com/companies/zapier" }, { "supports": "Listed as a Startup Weekend success story from Columbia", "url": "https://www.techstars.com/for/partners/startup-weekend" } ], "hot": 50, "id": "zapier", "key_idea": "Connect two apps that have no API integration by automating the steps a human would take in a browser.", "learn_from": "The Startup Weekend version was a demo of one integration. The product is the platform that lets anyone create the next thousand.", "links": { "official": "https://zapier.com" }, "name": "Zapier", "tags": [ "automation", "apis", "saas" ], "verified_at": "2026-10-24" }, { "body": "## What it is\n\nzkPow is a zero-knowledge system for proving work reputation without revealing the underlying data.\n\n## The problem\n\nA reputation score is only useful if it can be checked. Sharing the raw history to prove it exposes information the person may not want to disclose.\n\n## What makes it interesting\n\nIt won the ZK and Privacy category at ETH Porto 2023. The project separates the claim from the evidence, which is the core promise of a proof system.\n\n## What you can learn\n\nPrivacy is a feature of verification, not its opposite. A proof that reveals nothing can be more convincing than a document that reveals everything.\n\n## Limitations of this analysis\n\nThe record does not describe the circuit, the reputation model, or the cost of proving.\n", "description": "Zero-knowledge proof system for selective disclosure of work reputation without revealing underlying data.", "domain": "fintech", "event": { "award": "ZK/Privacy winner", "name": "ETH Porto", "year": 2023 }, "evidence": [ { "supports": "ETH Porto 2023 ZK/Privacy winner", "url": "https://www.recursive.so/hackathons" } ], "hot": 23, "id": "zkpow", "name": "zkPow", "tags": [ "web3", "zk", "privacy", "ethereum" ], "verified_at": "2026-10-24" } ] }