Y Combinator Startup Podcast

Poetiq is a new startup founded by former DeepMind researchers that recently achieved a major jump on the ARC-AGI and Humanity's Last Exam benchmark by layering a recursive self-improvement system on top of existing models. In this episode of Lightcone, Poetiq's Founder & CEO Ian Fischer joined us to discuss how small teams can build “reasoning harnesses” that outperform base models, what that means for startups and why automating prompt engineering may be one of the most powerful levers in AI today.Chapters:00:00 – Intro00:40 – What Is Poetiq?01:07 – Recursive Self-Improvement Explained02:07 – The Fine-Tuning Trap02:59 – “Stilts” for LLMs03:14 – Recursive Self-Improvement vs. Fine-Tuning05:05 – Taking the Top Spot on ARC-AGI06:37 – Beating Claude on Humanity’s Last Exam08:40 – How the Meta-System Works10:26 – Beyond RL: A New S-Curve11:32 – Automating Prompt Engineering13:37 – From 5% to 95% Performance14:50 – Early Access & Putting Your Agent on Stilts16:17 – From YC Founder to DeepMind Researcher18:29 – Advice for Engineers in the AI EraApply to Y Combinator: https://www.ycombinator.com/applyWork at a startup: https://www.ycombinator.com/jobs

With the takeoff of OpenClaw and MoltBook, a new agent-driven economy is taking shape.In this episode of the Lightcone, we took a look at the explosive growth of AI dev tools and whether the time has come for builders to make something agents want.

A very special guest on this episode of the Lightcone! Boris Cherny, the creator of Claude Code, sits down to share the incredible journey of developing one of the most transformative coding tools of the AI era.

In the AI era, startups aren't winning by hiring faster — they're winning by automating as many internal functions as possible. In this episode of Main Function, Garry breaks down how tiny teams are beating companies 20x their size by building automations into every workflow, from engineering to ops to customer support.

You’ve probably already heard all about OpenClaw (formerly Clawdbot/Moltbot). The viral sensation is an open-source AI assistant that runs on your own device, connects with messaging apps you already use, and goes beyond chat to actually execute tasks like managing your email, calendars, files, workflows, and more. Now meet the man behind it. YC’s Raphael Schaad sat down with Peter Steinberger, the creator of OpenClaw, to discuss the “aha” moment behind the viral personal AI agent, why local-first agents could replace many of today’s apps, and how personal agents will reshape the future of software.Chapters:00:00 – OpenClaw takes over the internet00:44 – Life after going viral01:28 – Why OpenClaw took off, what sets it apart02:56 – Bots talking to bots (and hiring humans)04:11 – From “God AI” to swarm intelligence05:07 – Peter’s original “aha” moment06:38 – Rebuilding the agent as a conversation07:38 – The moment it exceeded expectations10:21 – Are apps going to disappear?12:31 – Memory, data silos, and ownership14:39 – The privacy reality of personal agents15:05 – Letting the bot loose in public Discord16:55 – Giving an agent a personality18:19 – Contrarian building philosophy20:09 – CLIs vs MCPs21:28 – Building for humans first21:46 – The road aheadApply to Y Combinator: https://www.ycombinator.com/applyWork at a startup: https://www.ycombinator.com/jobs

Wondering why your maker-turned-manager suddenly seems distracted in meetings? Maybe they're addicted to coding agents! In this episode of Lightcone, Calvin French-Owen — a co-founder of Segment and former engineer on OpenAI's Codex team — joins us to talk about why coding agents suddenly feel so powerful, the differences between Codex, Claude Code, and Cursor, and what the future of work will look like.

When you're starting out, it isn’t enough to just build a minimum viable product. You also need a minimum evolvable product - one that can adapt to the needs of those critical early customers. In this episode of Main Function, YC General Partner Ankit Gupta offers an update to the classic MVP playbook. He’ll outline strategies for getting your first customers, the power of adaptability and how feedback from early users will ultimately shape the future of your product and your company.Apply to Y Combinator: https://www.ycombinator.com/applyWork at a startup: https://www.ycombinator.com/jobsChapters:00:00 – The Minimum Evolvable Product00:46 – Finding the First Believers01:29 – Counterintuitive Rules To Get Early Users02:10 – Learn Fast, Don’t Fear Churn02:52 – How Early Users Shape the Market You Enter04:22 – Tesla Case Study05:14 – How To Build To Evolve

Stoke Space is racing to build the world's first fully reusable rockets that can launch, survive reentry, and fly again and again. In this episode of Hard Tech, YC’s Aaron Epstein sits down with Stoke Space co-founders Andy Lapsa and Tom Feldman to find out why they chose to take on one of the hardest problems in rocket science, how an obsession with efficiency gives them an edge, and what full reusability could unlock for the future of spaceflight.Apply to Y Combinator: https://www.ycombinator.com/applyWork at a startup: https://www.ycombinator.com/jobs

2025 was the year AI stopped feeling chaotic and started feeling buildable. In this Lightcone episode, the YC partners break down the surprises of the year, from shifting model dominance to why the real opportunity is moving back to the application layer, and why the next wave of AI startups may be just getting started.

ARC-AGI is redefining how to measure progress on the path to AGI - focusing on reasoning, generalization, and adaptability instead of memorization or scale.During this month's NeurIPS 2025 conference, YC's Diana Hu sat down with ARC Prize Foundation President Greg Kamradt to find out why most AI benchmarks fail, how ARC-AGI reveals the limits of today’s models, and why measuring intelligence may be harder than building it.

Head of Design Ryo Lu helped transform Cursor from a feature-layer on top of VS Code into one of the world's leading AI code editors.He joins YC's Aaron Epstein on Design Review to talk about the path that brought him to Cursor, how rapid prototyping reshaped the core product and how he's breaking down the barriers that once separated designers and coders.

In just a few years, James Hawkins took PostHog from an idea hacked together right before YC's W20 deadline to a unicorn powering product analytics for thousands of teams.He joins YC's Brad Flora to talk about surviving six months of "pivot hell," why open-source analytics was the breakthrough, and how PostHog grew from fighting for its first users to launching full product lines—plus what he's learned about momentum, staying close to customers, and using transparency and humor to build a company that stands out.

Every major shift in consumer tech has a moment when it suddenly becomes accessible to millions. Michael Mignano helped spark one of those moments with Anchor, making podcast creation something anyone could do with a tap. Now at Lightspeed, he sees AI bringing a similar leap to music, media, and everyday apps.In this conversation, he and Garry trace the arc from the early days of social audio to today's consumer AI boom—and dig into what founders should focus on as the next generation of creative tools takes shape.

Cursor Head of Design Ryo Lu has spent his career at the intersection of design and engineering—from building fan sites as a kid to designing products at Stripe, Asana, and Notion. Now he's rethinking how software itself gets made.On this episode of Design Review, Ryo joins YC's Aaron Epstein to break down how great product websites communicate what a company does. They walk through sites from early-stage startups, calling out the small choices in structure, clarity, and brand that help users understand a product instantly — and the ones that get in the way.

Starcloud recently made history by launching a satellite with an NVIDIA H100 into orbit — the first time a GPU that powerful has ever operated in space. It's the first step toward building AI data centers in orbit, powered by continuous sunlight and cooled by radiating heat into deep space.Their approach could one day rival the world's biggest data centers while using less energy, zero fresh water, and far lower emissions.In this episode of Hard Tech, YC's Aaron Epstein visits Starcloud's HQ, where co-founders Philip Johnston, Ezra Feilden, and Adi Oltean explain how they built a working prototype in just 15 months — and why big tech is racing to space for AI compute.

Most founders think hiring is about interviewing. But it's actually about selling.For Startup School, Juicebox co-founder & CEO David Paffenholz joins YC's Harj Taggar to share how early-stage founders can find, pitch, and close top engineering and sales talent— from crafting better outreach to winning great hires from Big Tech— even when you're an unknown startup.

Jake Heller is the co-founder & CEO of Casetext, the AI legal startup behind CoCounsel, which was acquired by Thomson Reuters for $650 million.In his talk at AI Startup School on June 17th, 2025, he shared how his team did it—from picking the right idea to building AI products that actually work—and how founders can turn a cool demo into a reliable tool used by real customers.

Nearly every modern AI model, from ChatGPT and Claude to Gemini and Grok, is built on the same foundation: the Transformer.In this video, YC's Ankit Gupta traces how AI learned to understand language — from early RNNs and LSTMs to attention mechanisms and the breakthrough 2017 paper Attention Is All You Need — the discovery that unlocked the modern AI era.

Every founder faces moments where they’re not sure what to do next — such as how to go to market with AI products, when to pivot, and who/when to hire.In this episode of Office Hours, YC partners Pete Koomen, Brad Flora, Nicolas Dessaigne, and Gustaf Alströmer answer real questions from founders and share stories about how great teams build conviction, learn faster, and make better decisions as they grow.

For years, we've heard two major narratives about AI. One predicting the end of human work, the other dismissing it as hype. The truth is more nuanced, and more hopeful.From radiology to software engineering, the pattern repeats: as technology makes tasks cheaper and faster, demand for human creativity and judgment grows.YC's Garry Tan explores what history, economics, and real companies show us— that technology doesn't replace people, it redefines what we can do.