The next generation of startups will be built by smaller teams than ever before. At Startup School 2026, YC's @garrytan explains why we're we're entering the era of personal AGI: AI agents that run on your own infrastructure, compound your knowledge over time, and dramatically increase your ability to build. He shares the tools and workflows he uses every day, why every founder should own their intelligence instead of renting it, and what it means to build under your own power. 00:07 — What Founders Can Learn From Spinoza 04:46 — Personal AGI Is Already Here 07:57 — Why AI Makes One Person More Powerful Than Ever 12:29 — Your Life Is a Library 15:15 — Inside My Personal AI System 17:06 — Markdown Is Code 18:20 — Latent Space vs. Deterministic Code 21:13 — Building a Company of One 24:16 — How to Build Your Own Personal AGI 27:58 — Why Most People Will Quit Too Soon 29:09 — Own Your Skills Before Someone Else Does 32:24 — Personal AGI Means Owning Your Intelligence 35:21 — Why I Open Sourced Everything 38:30 — A Personal AGI for One Small Boy 40:18 — It's All Made Up. You Get to Make It Up.

datacenters in space sound incredible, but when you breakdown the eng, they seem less impossible than simply ambitious the tradeoff: the sun is a giant fusion reactor that already exists with free abundant energy but you pay for expensive transport + difficult cooling/comms

By far the most _fun_ product in my group this batch. AgentFM gives developers an easy way to listen along to their agents as they chop away at their tasks and see what their teammates agents are up to.

Models aren't the bottleneck anymore, context is. gives agents live web context without developers having to build and maintain infra. Congrats @mynameisyahia on the launch 🚀

Intelligence cofounder @grx_xce wants to build an intelligence marketplace for AI capabilities, not just a leaderboard that tells you which model is best. Today, Design Arena users submit creative tasks and have multiple AI models attempt it. The company then collects users' preference about which result was best, which gives Intelligence data about which models perform well for particular kinds of work. Grace thinks that in the future, it will become way easier for companies and individuals to customize models for one narrow purpose and sell access to it on a marketplace. She explains:

We’re bringing predictive validity to drug discovery’s most unpredictable stage. Today we are introducing Olio Labs’ in vivo platform, designed to predict diverse human clinical outcomes from a single high-throughput experiment. (1/n)

after we got into yc, I remember our partner telling us straight up, "this is one of the worst ideas i've ever heard. but it's your company, so keep going if you want." it's a strange feeling being told your idea is bad while still being accepted into the batch. in hindsight, they were right, as we ended up pivoting through it all. massive respect for letting us find out for ourselves.

Today, we're launching Standard Machines @stanmachines. We build reinforcement-learning environments for chip design. We want to push frontier model capabilities forward, and enable small teams to tape-out advanced chips in months. Read more in our launch post,

Thanks so much for having me on The Lightcone podcast @ycombinator, @garrytan, @snowmaker, @sdianahu and @harjtaggar! Such a fun discussion about the origins of @Starcloud_ and data centers in space! 🙏😃

Can you beat frontier AIs at social strategy games? Today, we're releasing Multi-Agent Arena for free public access! Play games like Risk, Catan, or Poker against GPT, Claude, etc. We're using this run evaluations on human and model behavior in complex environments.

Today, we're launching Coarena by Coasty, an arena for models to compete in computer-use. Give Fable, Sol, gemini, etc, the same computer task and watch them battle, for free. You only find out who you voted for after you pick a winner.

This is a brilliant approach to missile/drone interceptors—including the idea of bypassing glacial military buyers and selling directly to private companies who are threatened by the Iranians. Wishing them all the success.

We just closed Q2, and July was our strongest quarter-end month at Warp Summer is the quiet stretch for payroll and HRIS. Most companies switch at year end, so the industry tends to coast through it. We used it to move faster. We booked 10x the net new business we did last July. Net revenue retention held at 125% as existing customers kept expanding onto more of the product. Payroll volume on the platform grew 20% in the month. Companies are running real head-to-head evaluations against the incumbents and choosing Warp. They're choosing it for compliance and AI automations that legacy players still solve (badly) by throwing people at the problem. Onwards

Today we’re officially launching on YC. The web was built for humans to browse, not agents to understand. We turn the live web into structured, continuously refreshed data for AI agents through one API. Trusted by 400+ customers.