YCombinator News on LinkedIn

Two Pakistani founders pitched to a room of Silicon Valley investors. They became the first Pakistani startup ever accepted into Y Combinator. Sidra Qasim and Waqas Ali grew up near Okara, Pakistan, and wanted to give traditional Pakistani leather artisans a way to sell their handmade shoes and accessories directly to the world, rather than through layers of exporters who kept most of the profit. They built Markhor around ethical, direct trade, the artisans who made each pair got fairly paid, and the brand told their story to customers abroad. They funded their first batch of shoes through a Kickstarter campaign that hit its goal within a day, then went further, in 2015, Markhor became the first Pakistani startup ever accepted into Y Combinator, Silicon Valley's most selective startup accelerator. The relationships they built there later helped launch their next company, Atoms, backed by investors including Garry Tan and Alexis Ohanian. You don't have to leave your roots behind to go global, sometimes your roots are exactly what makes you stand out. 👉 Tag a friend who still thinks "Made in Pakistan" can't compete on a world stage. by Zobia Malik #Pakistani #Entrepreneur #YCombinator #MadeInPakistan

This Y Combinator Summer'26 Batch company's Tagline is "AI Weapons of Mass Production" Isengard Industries wants allied governments to be able to manufacture precision-strike and counter-drone weapons themselves, on their own soil, instead of depending on a handful of defense primes. Founded by Francisco Serra-Martins (CEO) and James Tennant (CIO). Their company tagline is blunt: "Weapons of Mass Production" the arsenal-of-democracy idea, rebuilt for a war fought with drones. Francisco's background is about as directly relevant as it gets. He previously co-founded Terminal Autonomy, one of Ukraine's largest deep-strike drone companies, and scaled it to $60M in revenue with customers including the U.S. Army, NATO, and Ukrainian armed forces. Before that, he was a combat engineer in the Australian Defence Force and later ran Sonder Design, a Forbes-listed company. James spent a decade in private equity and private credit , including running his own fund and has led due diligence on defence-tech deals across Australia, Ukraine, and the Gulf. He's also completing a "PhD on economic warfare and hybrid threats". The core insight behind the company is a manufacturing one, not a technology one: it's not that nobody can build a strike drone or an interceptor anymore, plenty of companies can. The bottleneck is that most of that production still runs through a small number of expensive, slow-moving prime contractors. Isengard's bet is on distributed, sovereign manufacturing, helping allied nations build their own production lines instead of buying everything from the U.S. defence industrial base. This isn't a pre-revenue idea. Their systems are already fielded and combat-tested in Ukraine, they've closed a funding round backed by Backswing Ventures and others just this month, and they list contract engagement across Europe, the Indo-Pacific, the Middle East, and early conversations with the U.S. That's a genuinely unusual amount of real-world traction for a company this young. My take: on pure founder-market fit, this is about as strong a match as you'll find in the batch, someone who already scaled a $60M defence company handing distribution and manufacturing know-how to a finance operator who understands how these deals actually get structured. The bigger question is less "can they execute" and more "should this kind of capability scale this fast" and that's a judgment call to be honest.. Where do you personally draw the line on covering defence-tech startups in a series like this, is it fair game like any other sector, or does it deserve different treatment? #YCombinator #StartupReview #DefenceTech

Robinhood is taking another step toward bringing venture capital to retail investors. Robinhood Ventures Fund II is targeting a $200 million IPO at $25 per share, with plans to list on the NYSE on August 13. The fund will hold around 80 early stage private companies, with technology making up about 64.5% of the portfolio. A major focus is Y Combinator backed startups, with AI companies featuring heavily across the portfolio. Tasklet is currently the largest holding at 4.5%. But the opportunity comes with significant risk. RVII plans to use leverage equal to about 67% of total assets, while its investments in illiquid early stage companies could make its NAV and market price highly volatile. The bigger question is whether retail investors are ready for venture capital risk inside a publicly traded structure. Follow The StartUp Chronicle for more updates on artificial intelligence, Innovation, Jobs in Industry 4.0 and the global startup ecosystem. #Robinhood #VentureCapital #Startups #YCombinator #AI #Fintech #RetailInvestors #PrivateMarkets #Investing #Technology

What if learning about the stock market felt more like a game than a classroom? 📈🎮 Had the opportunity to attend an amazing session by OpenTrade, a Y Combinator-backed startup, at Hindu College. What I found most interesting was how OpenTrade is using games, challenges, market duels, and real market data to make financial learning more interactive and engaging for Gen Z. The session gave me a completely different perspective on how gamification can turn something as complex as the stock market into an experience that people actually want to participate in. Beyond the product, hearing about their journey and understanding how they think about building for Gen Z was equally insightful. Thank you to the OpenTrade (YC s26) team for an engaging and thought-provoking session! 🚀 #OpenTrade #YCombinator #HinduCollege #FinTech #GenZ #Finance #Gamification #StockMarket #Entrepreneurship

In a design file, the worst a collaborator can do is move a rectangle. In a live AI agent session, one of the participants can send the contract, issue the refund, or release the payment. That one difference is why Multiplayer AI is harder than Figma ever was. Y Combinator just put Multiplayer AI on their Request for Startups. The diagnosis is right. But the Figma comparison quietly hides something. Docs and Figma had to solve presence and shared state. Multiplayer AI has to solve a third thing neither ever needed. Shared control. Because when three people can drop into a live agent session and redirect it mid run, who is accountable for what it does next? An agent runs for nine hours. The person who started it is asleep. Someone changes the instruction at hour six. It sends something to a customer at hour nine. Now try to reconstruct who allowed that. This is why agent pilots look incredible in demos and never reach production. The capability was never the constraint. The container was. I wrote a full breakdown of what we learned building this at Softworker AI, including why half a governance model is not half as safe, it is differently unsafe. And one more thing. We are taking 10 teams as founding design partners. Full platform access, free, for all of early access. If your team runs agents in production and stalled at the approval problem, comment below or DM me. Full article is here Post 180/1000 #MultiplayerAI #AIAgents #SoftworkerAI #AIGovernance #BuildingInPublic #YCombinator

Amateurs talk strategy. Professionals talk logistics. When startups fail, especially in deep tech and hardware, it’s rarely because the underlying physics or technology didn't work. They fail because the human organization and infrastructure around the technology collapsed under its own weight. Max Hodak (CEO of Science, co-founder of Neuralink) recently shared insights on why operational infrastructure is the true bottleneck to startup speed: 💡 1. Speed is determined by infrastructure If you learn something every week and your competitor learns something every month, they will never catch up. Compounding speed isn't about working harder; it's about building systems (purchasing, budgeting, inventory) that eliminate friction for high-performing teams. 🛒 2. Purchasing friction kills momentum Handing out credit cards without budgets leads to micromanaging $3,000 power supplies. But putting complex B2B procurement in front of top engineers turns them into admin workers waiting weeks for parts. You need streamlined, internal systems to calculate true experiment costs without slowing people down. 👥 3. "EigenReviews" & continuous feedback Annual 360-degree performance reviews are often disruptive and tell you what you already knew. Instead, use continuous, graph-based feedback: ask team members every few weeks, "Knowing how this person turned out, would you vote again today for their hire?" Weighing feedback using graph centrality (similar to PageRank) surfaces performance trends without traumatic once-a-year HR cycles. ⚡ 4. Action produces information When stuck in a local minimum, static analysis won't save you. In physics, exerting action creates entropy and information. The action space is always larger than it appears—start taking steps to unblock the team and generate new data. 🎯 5. You cannot delegate your judgment Being average or copying consensus in startups means failure. The most important high-stakes decisions will happen when there is no one left to give you advice. You must cultivate and trust differentiated judgment when you are completely alone in a realization. Full video / transcript details: Max Hodak: Average Is Not Good Enough (Y Combinator Startup School). Watch here: https://lnkd.in/edB2tDH6 #Startups #DeepTech #Leadership #Biotech #Operations #Engineering #YCombinator #Neuralink

Y Combinator startups. Do you need to furnish your onsite HQ with Exec furniture? DM me. I am the startup furniture plug. I got several sets of Metro Office Tables. #ycombinator #startups

The most useful thing I did in my co-founder search was close LinkedIn. I started where every founder starts: searching on Linkedin. It got me nowhere, and here is why. Nobody at a big corporate posts “open to building something risky.” People between jobs rarely update their profile. And anyone willing to build on nights and weekends? Not listed either. Then I found out about Y Combinator Co-Founder Matching. What worked: 1. Spend real time on the filters. That is 80% of the outcome. 2. Review around 100 profiles before contacting anyone, so you know what a strong fit looks like. 3. Shortlist, then reach out. 4. Talk to one person at a time. Parallel conversations feel productive and teach you nothing. 3 months later, I have my co-founder finally. Thank you Y Combinator for building this platform. If you are a solo founder stuck at this stage: go where people have already raised their hand. #CoFounder #YCombinator #StartupLife #Founders #SoloFounder #Entrepreneurship #MedTech #BuildInPublic

I left YC Startup School 2026 with a slightly different view of what “moving fast” actually means. It’s easy to hear “build faster” and interpret it as shipping more. But after two days of listening to people like Jensen Huang, Sam Altman, Boris Cherny, Alexandr Wang, Dmitri Dolgov and others, I think it’s closer to this: learn faster, update faster, and don’t get too attached to what worked yesterday. A few things that stuck with me: → Jensen Huang spoke about resilience as a willingness to confront reality, learn, and keep going. → Boris Cherny shared that Anthropic removed over 80% of Claude Code’s system prompt as the models improved. That one really stayed with me — sometimes the thing you spent weeks engineering eventually becomes unnecessary. → Dmitri Dolgov described the demo as the first 1%. Getting something to work is one thing. Making it reliable enough for the real world is the actual game. → Alexandr Wang talked about developing your own view of the future instead of simply following consensus. And across almost every talk, there was the same underlying energy: start before you feel ready. What I enjoyed most, though, wasn’t just seeing these people on stage. It was being surrounded by thousands of builders who were all trying things, changing their minds, meeting people, and getting back to work. I’m leaving San Francisco with a lot more conviction around what I’m building with Tutix.ai, and an even bigger reminder that in AI, the goal isn’t to build something because the technology is impressive. It’s to use the technology to make something possible that wasn’t possible before. Very grateful to Y Combinator for putting together such an incredible two days, and to everyone I met along the way. Back to building. #YCStartupSchool2026 #YCombinator #AI #Startups #Building

OpenAI just cut GPT-5.6 Luna's price by 80%. Anthropic just published research where Claude found real weaknesses in cryptographic systems. One week, two bets on what actually wins the AI race, being the cheapest or being the sharpest. 6 things from this week in AI/startups most people missed 👇 1️⃣ OpenAI dropped GPT-5.6 Luna's input/output token prices by 80%, with Terra getting a 20% cut too. Undercutting your own newest model this hard is a distribution play, not a cost-saving one. It's a bet that usage volume matters more than per-token margin right now. 2️⃣ Anthropic published research showing Claude uncovered genuine weaknesses in cryptographic systems, using an internal tool called Claude Mythos Preview. That's a meaningfully different flex than a benchmark score, it's a model doing security research a human team would usually run. 3️⃣ NVIDIA and Firebird AI launched the CIS region's largest AI factory in Armenia, with Jensen Huang on hand for the milestone. Built with Dell, Schneider Electric, and Vertiv, it's aimed at giving Armenia and Kazakhstan their own AI infrastructure instead of renting compute elsewhere. 4️⃣ Harsha Gaddipati, CEO of Slashy, launched an "AI native" email platform that triages your inbox and drafts replies in your voice through an MCP-connected assistant. Inbox AI has been tried a dozen ways, the bet here is that voice-matching plus automation is the part that finally sticks. 5️⃣ Y Combinator Whatnot raised $545M in Series G funding at a $20B valuation to grow its live-commerce platform. It already blew past its entire 2025 sales volume by mid-2026, live shopping is quietly becoming a real distribution channel, not a novelty. 6️⃣ Gabriel Petersson, ex-OpenAI and Midjourney researcher, left OpenAI three months ago and launched Energy, a browser-based AI agent that turns multi-step web tasks into a single sentence. Another signal that some of the best research talent would rather build the agent layer than stay inside a frontier lab. My biggest takeaway is that the frontier labs are competing on completely different axes now, not just model size. Who owns the cheapest intelligence. Who owns the proof that it actually works. Who owns the infrastructure it runs on. Which one of these bets you think pays off first? #AI #OpenAI #Anthropic #NVIDIA #YCombinator #Startups Sources 1️⃣ https://lnkd.in/gjnQepkw 2️⃣ https://lnkd.in/g2dbVrMk 3️⃣ https://lnkd.in/g5W_WW9H 4️⃣ https://lnkd.in/dMmTQmHg 5️⃣ https://lnkd.in/gcSKyANi 6️⃣ https://lnkd.in/g28wKUCY

Y Combinator is betting on a new generation of startups. 🚀 The startup world is changing fast — and the latest move from Y Combinator is another signal of where the next wave of innovation could come from. From AI to new-age technology and ambitious founders, the opportunity is no longer limited to traditional startup ideas. The bigger question is: What will the next generation of billion-dollar companies look like? At The Idea School, we break down the ideas, businesses and trends shaping what comes next. Think better. Build smarter. Create impact. 🌐 theideaschool.in #TheIdeaSchool #YCombinator #Startups #AI #Entrepreneurship #Business #Innovation #StartupNews #Tech #FutureOfBusiness #Founders #VentureCapital

🚨 We're giving out $1,000. Backed by Y Combinator. We're running the OpenTrade (YC s26) Creator Challenge - and we want the people who can actually stop the scroll. 📝One Linkedin Post. No required format. No follower minimum. No previous experience required. - $500 to the best post. $250 for runner-up. $250 split across the next 5 strongest entries. - Every strong entry gets reshared by our page. - Standouts get invited to work with us directly as a Growth Fellow - paid, ongoing. We want the post that makes a total stranger actually stop scrolling. No follower minimum. No experience required. Just a real take. Entries close August 25! 👇 Comment “OpenTrade” + connect w/ me, Repost for higher chances! I'll send you the full rules + how to enter. Dylan Pak, Tario You, Ojas Kandhare #YCombinator #StartupLife #BuildInPublic #CreatorChallenge #StockComp

Y Combinator Startup School 2026 was a very cool experience as a first-timer — exceptional people, great talks, YCML (video below!), and saw Sam IRL :) 𝗬𝗖𝗠𝗟, 𝗬𝗖’𝘀 𝗶𝗻𝗮𝘂𝗴𝘂𝗿𝗮𝗹 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝘀𝘆𝗺𝗽𝗼𝘀𝗶𝘂𝗺, was undoubtedly my favorite part. Presented a poster, chatted about my work with Ankit, and experienced something I have never in my past conferences. My reflections on why 𝗬𝗖𝗠𝗟 has a unique place in the ecosystem. 𝟭. 𝗬𝗖 𝗮𝘁𝘁𝗿𝗮𝗰𝘁𝘀 𝗮𝗻 𝘂𝗻𝘂𝘀𝘂𝗮𝗹𝗹𝘆 𝗱𝗶𝘃𝗲𝗿𝘀𝗲 𝗰𝗿𝗼𝘄𝗱 I was talking about agents with high schoolers, AI researchers, biologists, people in construction, and founders across industries. What conference can match that mix? 𝟮. 𝗕𝘂𝗶𝗹𝗱𝗲𝗿𝘀 𝗺𝗲𝗲𝘁 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵𝗲𝗿𝘀, 𝗻𝗮𝘁𝘂𝗿𝗮𝗹𝗹𝘆 Many events try to manufacture this collision. At YCML, it already exists. The builders are there. The researchers are there. 𝟯. 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵-𝘁𝗼-𝗺𝗮𝗿𝗸𝗲𝘁, 𝗮𝗰𝗰𝗲𝗹𝗲𝗿𝗮𝘁𝗲𝗱 Universities have long played this role, and conferences are increasingly leaning into it too. I first saw this at ECCV ’22, hearing Michael Black. But 𝗬𝗖𝗠𝗟 pushes it much further. Imagine being in one room with a potential co-founder, investor, or even your first customer. Very positive about this, and so much looking forward to the next one! Link to Youtube: https://lnkd.in/g7RV2ChA Link to X: https://lnkd.in/gwkqPKtg #YCombinator #YCML #YC #LLM #VLM #Agents #AI #Research #YCSUS #GenAI #Evals #Benchmarks

Praxis (YC S26) launched out of Y Combinator paying businesses to record their staff doing physical work the actual motion, captured on headsets and wrist cameras to train robotics models. The line that matters isn't in the hardware spec. It's the qualifying criteria: there isn't any. Local businesses through to conglomerates worth 3% of a nation's GDP. If physical work happens on your premises, you're a potential supplier. Here's what makes this interesting to us as a studio. A new asset just became sellable, and none of the infrastructure around it exists. No going rate. No standard contract. No benchmark to check an offer against. No advisor who has done this ten times before. That's not a criticism of Praxis that's what every new market looks like on day one. Demand arrives first. Everything that makes the market fair and liquid gets built afterward, usually by other people. Those other people are typically where the second wave of value sits. For operators, the practical version is simpler: you now own something with a buyer, and no reliable way to price it. That's a bad position to negotiate from and a worse one to be unaware of. We wouldn't tell every firm to sell. You'd be licensing a recording of work a model may eventually do. But every firm should have a view, because the first offers are already going out. Congrats on the launch Praxis (YC S26)! A genuinely ambitious place to start. Optivance Advancements #VentureStudio #VentureBuilding #Startups #Founders #Entrepreneurship #BusinessStrategy #OptivanceAdvancements #VentureCapital #Ycombinator #YC #Robotics #Data

Let's get you hired asap. 10 fresh jobs from Y Combinator startups: Senior Counsel at Method Financial (United States) https://lnkd.in/g_gjPzgs Corporate Attorney at Corgi Insurance (United States) https://lnkd.in/gm5hD-Ph Founding Account Executive at Moritz (United States) https://lnkd.in/gs4q6QHt Mechanical Engineering Placement (6 Months), New Grad at Aseon Labs (United States) https://lnkd.in/gsQ5gM7N Founding Design Engineer at Moritz (United States) https://lnkd.in/g-zadXes Founding Electrical Engineer (Full-Time) at Aseon Labs (United States) https://lnkd.in/gGgcDV6j Compliance Officer, Options Principal (Series 4) at Hapi (Remote) https://lnkd.in/gDQ-5Ehz Founding GTM at Ventura (United States) https://lnkd.in/g5TRTkmz Human Resources (HR) Intern (Paid - India/Remote) at Peakflo (India / Remote (India)) https://lnkd.in/gQfbFFfS Product Manager at dili (United States) https://lnkd.in/g6TrgMig More YC roles live now at https://lnkd.in/enFdH6i4 #ycombinator #startups #hiring #jobs #techjobs

One company. Ten open roles. Here's the story of each one. They hit 8 figures in ARR in year one. Then doubled it in year two. 150K+ GitHub stars. Series A backed by Y Combinator, Zapier, and Shopify's Tobias Lütke. They build the infrastructure other AI companies run on. And right now, they're hiring for almost everything. ⚙️ Product Engineer — $210K–$260K + Equity The person who owns how the core product works for developers. Your name is on the reliability, the response format, the latency, and every edge case. No PM between you and the problem. 💼 Revenue Operations Lead — $190K–$240K + Equity The one who finds what's broken before anyone notices. Billing, dashboards, compliance, procurement — you own the operational backbone of a rocketship. 📈 Head of Growth Marketing — $200K–$250K + Equity Owns the number. Not the strategy deck — the actual weekly signup pacing across SEO, content, partnerships, and channels that compound. Developer audience. Real accountability. 🔧 Sales Engineer — $180K–$240K base · OTE + Equity The person developers trust because you earned it. You write the POC, answer the hard question, and move the deal. Technical enough to be credible. Commercial enough to matter. ⚖️ In-House Counsel — $220K–$265K + Equity First lawyer in the building. Web scraping, data licensing, public data law — a genuinely novel legal frontier. Founder access. Real authority. 🖥️ Backend Infrastructure Engineer — $200K–$250K + Equity The one who keeps success rates high against sites that actively try to stop you. Proxy systems, throughput, adversarial reliability. Hard problems with no playbook. 🗺️ Head of Product — $250K–$290K + Equity First proper Head of Product. Reports to the CEO. Owns the roadmap and the hard prioritization calls at one of the fastest-growing AI infrastructure companies in the world. 🔗 SEO Specialist — $135K–$180K + Equity First dedicated owner of off-page authority. Link acquisition, digital PR, and figuring out how AI-driven discovery is changing organic growth for a developer tool. 🔍 Search Engineer — $190K–$260K + Equity Build the systems that let anyone turn the open web into a search index. Crawling, indexing, retrieval at scale. Billions of docs. Real hard infrastructure problems. ✍️ Technical Writer — $140K–$185K + Equity Docs that turn "what does this do?" into "I shipped it in ten minutes." First dedicated writer at a company where developer experience is everything. All roles are hybrid in San Francisco [Relocation package available for the right candidate]. All are confidential until the right conversation. Apply directly or refer someone 👇 jobs.trustyrecruit.com #NowHiring #SanFrancisco #TechJobs #AIJobs #StartupJobs #YCombinator #ProductEngineer #HeadOfProduct #InHouseCounsel #SalesEngineer #GrowthMarketing #TechnicalWriter #BackendEngineering #SearchEngineering #SEO #RevOps #TrustyRecruit

Didn't exactly expect a word guessing game at a college seminar to end with an internship offer. 😄 I recently attended a seminar organised by Kaizen- Department of Commerce, Hindu College, in collaboration with OpenTrade (YC s26), a Y Combinator backed startup building in the investing space. It was easily one of the most fun and interactive sessions I've been part of. The highlight was Tickerdle - OpenTrade (YC s26)'s ticker guessing game, where participants identified companies based on the number of blanks and clues given. Correct answers won cash prizes of up to ₹10,000, which made the room incredibly competitive and engaging. And somehow, doing well at Tickerdle turned into a very unexpected outcome - an internship offer. Definitely not what I had in mind walking in, which is probably what made it so memorable. Beyond the game, it was great getting to learn more about what the OpenTrade (YC s26) team is building, and their approach to making investing more engaging for a new generation of investors. A big thank you to Suraj Mishra, Dylan Pak, Tario You and @ojaskandhare for making the session such a blast — and for the unexpected opportunity. And kudos to Kaizen Hindu for organising such a refreshing, memorable event on campus. Excited for what's ahead! 🚀 #OpenTrade #YCombinator #HinduCollege #KaizenHindu #FinTech #Investing #Startups

Y Combinator publishes a list of the startups it wants to fund. Most founders skim it. It's actually the clearest signal in venture. The Fall 2026 edition just landed — 13 requests. The throughline: AI is done optimizing software and has moved into the physical world. A few that stood out: → Compute at Sea. AI is running out of electricity and land. YC's answer: floating data centers. 70% of the planet is ocean — no permitting, natural cooling. → Proving You're Human. A finance worker joined a video call with his CFO and wired out $25M. Every other person on that call was a deepfake. Every trust signal we have was built for a world where faking a human was expensive. That world is gone. → The Future of American Defense. Written by the sitting U.S. Secretary of the Army — the first time a government official has authored an RFS. → New Operating Systems for the Physical World. 80% of the global workforce doesn't sit at a desk, and their software hasn't changed in 20 years. Now there are three kinds of workers to coordinate: agents, robots, and humans. → The Best Time to Build in Crypto. Prices are down, builders are leaving, and YC says that's exactly the point — bear markets attract founders who build instead of exit. The pattern worth noticing: nine of the thirteen require you to touch something real. Hardware, regulation, physical infrastructure, or all three. The pure-software land grab is getting crowded; the moat has moved to the messy parts. Full list: ycombinator.com/rfs Which one is most underrated? #ycombinator #startupideas #buildinpublic #aistartups #founders #venturecapital #startupfounder #ai #entrepreneurship #techstartups #startupdubai #uaestartups

The Founders Who Build Iconic Companies Start With Ideas Everyone Else Dismisses Alexandr Wang, founder of Scale AI and now Chief AI Officer at Meta, shared a simple but powerful truth at Y Combinator ’s Startup School 2026: If everyone already agrees your idea is great, someone may have already built it or the opportunity may have passed. The best founders often start with ideas that seem strange, small, unconventional, or even wrong to most people. That is precisely where extraordinary opportunities can hide. Paul Graham has made a similar point for years: most great startup ideas look like wells, not craters. You begin with something a small number of people desperately need, rather than something a large number of people merely like. Facebook began exclusively for Harvard students. Microsoft began with Altair BASIC, created for a relatively small community of hobbyists. What appears insignificant to everyone else can become the foundation of something enormous. Wang also offered an important reminder to aspiring founders: No one is naturally good at starting a company. Successful entrepreneurs become better by building, making mistakes, learning, and continuing to believe in an idea long before the market catches up. Developing your own internal compass for how the future will unfold and maintaining conviction despite the noise is one of the most important skills a founder can develop. The real danger isn't building something too small. The danger is building something too safe. If your idea doesn't make you feel at least a little uncomfortable, it may already be taken. Think differently. Start small. Build with conviction. #AlexandrWang #YCombinator #StartupSchool2026 #FounderWisdom #PaulGraham #StartupIdeas #Conviction #Entrepreneurship #AI #ScaleAI #Meta #Innovation #StartupAdvice #ZeroToOne #StartupMindset #Africanstartupmetrics

I WANT TO SLEEP !!!!!!!!!!!! Over the last 3 days, I've been building something that's consumed almost every hour of my day. Not another AI wrapper. Not another weekend project. And definitely not something I want to reveal just yet. It's the result of everything I've learned over the last 1.5 years of coding. Every mistake. Every failed side project. Every late night debugging session. Every open source rabbit hole. Every product teardown. Every conversation with founders. I'm compressing all of that into one product. The market is changing fast. AI is rewriting how software is built, open source is evolving at an incredible pace, and the bar for building meaningful products keeps getting higher. Instead of chasing every trend, I'm trying to build something that could genuinely matter. If anyone from Y Combinator , Y Combinator Community , Antler, or any startup accelerator comes across this, I'd love your perspective. A few questions I'm thinking about: • What makes a first-time founder stand out today? • What convinces you that a product is worth backing before it has massive traction? • What mistakes do technical founders consistently make in their first serious product? I'm not looking for shortcuts. I'm looking for honest feedback from people who've seen thousands of founders before me. Still building. More soon. #BuildInPublic #Startup #AI #OpenSource #SaaS #Founders #Entrepreneurship #IndieHackers #SoftwareEngineering #ProductDevelopment #Developer #Innovation #YCombinator #Antler #TechStartups