A 7-million parameter model outperforming models a thousand times its size on tasks like ARC Prize. That's what recursive reasoning unlocks.In this episode of Decoded, YC's Ankit Gupta and Francois Chaubard break down two recent papers on recursive AI models, HRMs and TRMs, that are achieving state-of-the-art results with a fraction of the parameters of today's largest models.They explain why standard LLMs hit a fundamental ceiling on certain reasoning tasks, how recursion at inference time gives small models the compute depth to break through it, and what happens when you combine these ideas with the power of large-scale foundation models.

Demis Hassabis has had one of the most extraordinary careers in tech. He started as a chess prodigy and video game designer at 17 before getting a PhD in neuroscience and going on to found DeepMind. His lab cracked Go, solved protein structure prediction with AlphaFold, and then gave it away free to every scientist on earth. That work won him the 2024 Nobel Prize in Chemistry. Today he leads Google DeepMind, pushing toward the same goal he set as a teenager: AGI. On this special live episode of How to Build the Future, he sat down with YC's Garry Tan to talk about what still needs to happen to get us to AGI, his advice for founders on how to stay ahead of the curve and what the next big scientific breakthroughs might be. Chapters:00:00 — Intro00:46 — Demis Hassabis: From Chess Prodigy to DeepMind01:48 — What’s Missing Before We Get To AGI?03:36 — Why Memory Is Still Unsolved06:14 — How AlphaGo Shaped Gemini08:06 — Why Smaller Models Are Getting So Powerful10:46 — The 1000x Engineer12:40 — Continual Learning and the Future of Agents13:32 — Why AI Still Fails at Basic Reasoning15:33 — Are Agents Overhyped or Just Getting Started?18:31 — Can AI Become Truly Creative?20:26 — Open Models, Gemma, and Local AI22:26 — Why Gemini Was Built Multimodal24:08 — What Happens When Inference Gets Cheap?25:24 — From AlphaFold to the Virtual Cells28:24 — AI as the Ultimate Tool for Science30:43 — Advice for Founders33:30 — The AlphaFold Breakthrough Pattern35:20 — Can AI Make Real Scientific Discoveries?37:59 — What to Build Before AGI ArrivesApply to Y Combinator: https://www.ycombinator.com/applyWork at a startup: https://www.ycombinator.com/jobs

Replit is the leading no-code app builder for consumers and enterprise, letting anyone with an idea build real, deployed software using natural language. The company just raised a $400 million Series D at a $9 billion valuation.In this episode of Founder Firesides, co-founder and CEO Amjad Masad sat down with YC's Andrew Miklas to talk about Replit's 10-year journey from browser IDE to vibe coding platform, why the people getting the most value aren't traditional developers but founders and domain experts closest to the problem, and what Agent 4 unlocks with parallel agents, built-in design, and the ability to run your entire company on Replit.

AI isn't just making teams more productive. It's changing how companies should be built.

In this episode of Startup School, YC Partner Diana Hu explains what it means to build an AI-native company, where AI isn't just a tool but the operating system your company runs on.

She breaks down how to make your company queryable so agents can improve across every function, why management hierarchies break down when an intelligence layer replaces human middleware, and why early-stage founders have a massive edge in building this way from day one.

Even the most successful websites eventually need a redesign. Take Stripe for example. After six years with the same homepage, they recently unveiled a brand new site that reflects how the fintech giant has evolved over the past few years. So when is the time right for a new landing page? And what should you prioritize in the redesign? In this episode of Design Review YC’s Aaron Epstein sat down Stripe’s Head of Design Katie Dill to pull the curtain back on their high profile redesign and to discuss how their team is evolving in a world dominated by new AI design tools.

Physical Intelligence is building a foundation model that can control any robot to do any task — what the team describes as the GPT-1 moment for robotics.

The company's cross-embodiment approach trains across many different robot platforms, and recent results show tasks being performed zero-shot that last year required hundreds of hours of data collection.

In this episode of The Lightcone, co-founder Quan Vuong sat down with Garry, Jared, Diana, and Harj to talk about why robotics is finally ready for its scaling moment, how PI runs its models in the cloud rather than on-device, and the playbook for what Quan sees as a Cambrian explosion of vertical robotics companies.

1 in 11 babies born in America this year will be screened by a genetic test that didn't exist a decade ago.Biotech startup BillionToOne turned a simple but radical idea—detecting rare fragments of fetal DNA in a mother's blood—into one of the most widely used prenatal tests in the U.S. And they're not stopping there. The same approach could unlock something even bigger: early-stage cancer detection from a blood test, a breakthrough that could one day save millions of lives.In this episode of Hard Tech, YC's Jared Friedman sits down with David Tsao and Oguzhan Atay to hear how they went from half a lab bench to a $4B biotech company—and why they believe this is just the beginning of what their technology can do.

In this episode of Founder Firesides, YC Managing Partner Jared Friedman talks to Karine Mellata, co-founder of Variance (W23), who is coming out of stealth and announcing their $21 million Series A.

Variance builds purpose-built AI agents for risk and compliance — automating fraud detection, content review, and identity verification for Fortune 500 companies and platforms like GoFundMe. They discuss why Variance built in the shadows for three years, detecting state-sponsored fraud rings, and the accident that nearly ended the company.

François Chollet has spent years asking a different question than most of the AI world. Instead of scaling what already works, he’s trying to understand what intelligence actually is—and how to build it from first principles. In this episode of Lightcone, he traces that path from his early work on deep learning to the creation of the ARC prize, and the launch of ARC V3, a new benchmark designed to measure something deeper than performance: the ability to learn, adapt, and reason efficiently in entirely new environments. He explains why today’s systems may be hitting limits, what recent breakthroughs really mean, and why reaching true general intelligence may require a fundamentally different approach.00:00 - AGI by 2030?00:31 - Introducing Ndea: A New Path Beyond Deep Learning01:08 - A New ML Paradigm 01:30 - Replacing neural nets with compact symbolic programs03:04 - Why Ndea Isn’t Competing With Coding Agents05:20 - Why Everyone Might Be Wrong About Scaling LLMs07:22 - Why Coding Agents Suddenly Work So Well08:50 - The Limits of LLMs in Non-Verifiable Domains10:48 - What AGI Actually Means (And Why Most Definitions Are Wrong)13:30 - Why Deep Learning Hits a Wall 14:00 - ARC’s Origin Story18:20 - ARC Benchmarks Explained: From V1 to V322:49 - The RL Loop Powering Coding Agents Today27:03 - ARC-AGI V3: Measuring “Agentic Intelligence”31:14 - Inside the ARC Game Studio35:31 - Could AGI Fit in 10,000 Lines of Code?44:01 - Building Ndea: From Idea to Compounding Research Stack46:46 - The Future of ARC: Benchmarks That Evolve With AI47:21 - Why There’s Still Huge Opportunity for New AI Paradigms53:37 - How to Build a Breakout Open Source Project - Lessons From Kera56:39 - Advice For How To Think About AIApply to Y Combinator: https://www.ycombinator.com/applyWork at a startup: https://www.ycombinator.com/jobs

Solugen is reinventing the trillion-dollar chemical manufacturing industry by combining biology and chemistry in a new way. In this episode of Hard Tech, YC's Jared Friedman visits co-founders Gaurab Chakrabarti and Sean Hunt at their Houston HQ to see how they went from a $7,000 PVC reactor to a billion-dollar company competing with industry giants. They cover the breakthrough behind their enzymatic + catalytic production, how they found their first customers, and why starting small and staying close to customers let them win in a capital-intensive industry.

Chapters:00:00 - A New Kind of Chemical Plant01:02 - Fusing Biology + Chemistry In a New Way02:23 - The Eureka Moment: From Pancreatic Cancer to Hydrogen Peroxide03:30 - Using A Sugar Feedstock Over Oil and Gas 04:22 - Proving Enzymes Work at Scale In Chemical Manufacturing05:16 - The $7K PVC Reactor06:44 - Finding First Customers at YC08:12 - What The Co-founders Got Out of YC09:33 - Seed Round to Bio Forge10:32 - Scaling to a Full-Size Plant (Bioforge)11:57 - The Future of American Manufacturing12:29 - The Next Decade of Solugen

Apply to Y Combinator: https://www.ycombinator.com/applyWork at a startup: https://www.ycombinator.com/jobs

In this episode of The Lightcone, we talk with Mukund and Madhav Jha, the founders of Emergent - an AI platform that lets anyone build and ship production-ready software. In just eight months, users have created more than 7 million apps on Emergent, with the number doubling in just the last 45 days. We discuss how they built one of the most powerful AI coding agents, why they focused on non-technical users and what it's like building in India for a global audience.

Apply to Y Combinator: https://www.ycombinator.com/apply

Chapters:

00:00 - Intro

01:06 - What Is Emergent?

01:18 - Founder Backstory

02:09 - From AI Testing to General Coding Agents

02:52 - Getting Ahead of the Market

04:18 - The Pivot to Non-Technical Users

05:22 - Why Second Movers Can Win in AI

09:04 - Building for Production, Not Just Prototypes

18:21 - Live Demo: Building Apps with Emergent

24:40 - How Emergent Hires and Runs a Lean Team

29:04 - Is SaaS Dead? The Rise of Personalized Software

34:04 - The Future: Niche Apps, Solo Builders and AI Agency