The Lobster Talks Podcast by Lobster Capital

Most founders think they need to monetize on day one, but this YC breakout grew to 2 million weekly downloads by giving everything away for free, and now they power ChatGPT. In this episode of The Lobster Talks, we sit down with Bereket, the founder of BetterAuth, to dissect the ultimate open-source distribution playbook. Moving directly from Ethiopia to Silicon Valley, Bereket built an authentication powerhouse that is disrupting legacy vendors and establishing the new standard for the AI agent economy. We break down the transition from human auth to agent identity, the playbook behind corporate acquisitions like Anthropic buying Bun, and why rushing your monetization strategy can kill your hypergrowth. What you will learn: The exact mechanics BetterAuth used to scale from 150K to 2 million weekly downloads without a marketing budget.Why building a high-quality open-source ecosystem is the highest-leverage distribution hack in tech.How Anthropic’s acquisition of Bun proves that massive developer utility trumps early monetization.The unbundled future of identity: Why the massive explosion of AI agents requires a completely new auth protocol.How Claude Code and LLM recommendations are completely replacing traditional B2B SaaS marketing funnels. 📌 Watch now and subscribe for more smart, fast, founder-first interviews. Chapters: 00:00 Open Source Distribution 00:27 YC Roaster Sponsor 01:07 Meet BetterAuth Founder 01:36 BetterAuth Explained 02:21 Insane Growth Story 03:28 How Word of Mouth Wins 06:26 Open Source Playbook 08:50 Monetizing Open Source 09:58 Docker and Bun Lessons 14:45 Avoiding Hosted Auth Trap 15:44 Agent Auth Breakthrough 17:43 Protocol and Future Plans 19:49 Why Not Hosted Yet 20:54 Bottom Up Standards 21:42 Mission For Agent Auth 23:06 Ethiopia To YC Journey 24:33 YC Content Flywheel 25:45 Family Reactions And Impact 27:55 Better Auth Roadmap 29:16 Better X Side Projects 30:23 AI Coding And Momentum 34:16 Winning AI Recommendations 36:19 Final Advice And Wrap 🎧 Listen on the go: Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP Apple: https://apple.co/4cZ8RMm My other YouTube Channel: https://www.youtube.com/@lobstercap This Lobster Talks episode is brought to you by YC Roaster, a free tool helping founders improve their YC applications. 🔗 https://ycroaster.com

YC just added one question that quietly changes what it means to be a founder. And it’s not about your TAM. It’s not about your pitch. It’s about how you use AI to build. In this episode of Lobster Talks, we break down YC’s newest application change, the rise of coding agents like Claude Code, and what this signals about where startup building, and venture capital is heading next. We also go deep on YC Roaster, Lobster Capital’s new initiative helping founders improve their YC applications using AI + YC alumni review. Here’s what you’ll learn: - Why YC’s new “coding agent session” question is a founder filter - How vibe coding is becoming table stakes in 2026 - What 20,000 YC applications really signal about global founders - Why traction keeps rising as the YC acceptance bar - How smart VCs are building internal tools and why most aren’t - Why the best founders and funds pivot fast This is a behind-the-scenes look at how YC evaluates builders today. 📌 Watch now and subscribe for more smart, fast, founder-first interviews. Chapters: 00:00 Claude Code Is Addictive: Why Everyone’s Vibe Coding Now 00:44 Sponsor Break: YC Roaster - Free, Brutally Honest YC Application Feedback 01:27 Welcome to Lobster Talks: What’s New in the YC World 02:00 How YC Roaster Was Built (Fast) + How the AI→Alumni Review Flow Works 03:37 Launch Results: Hundreds of Apps, MVP Bugs, and the V2 Relaunch Plan 06:23 What Applicants Look Like: Quality Spread, Global Submissions, and Reviewer Time Sinks 11:29 Why Build YC Roaster: Lobster’s “Helpful Fund” Flywheel + VCs Who Ship 14:20 YC’s New Application Question: Submit a Coding Agent Session You’re Proud Of 15:06 What YC Is Really Testing: Technical Bar, AI Fluency, and Advanced Agent Workflows 22:42 Founder vs Idea: Pivot Culture, Proof-of-Work, and How AI Changes Evaluation 27:06 VCs Need Tools Too: Internal Engineering, Interactive Marketing, and Fund Differentiation 29:28 YC Request for Startups: 2026 Themes - Agents, Stablecoins, GovTech, Spatial/Robotics 31:33 Wrap-Up: Subscribe, Reviews, and Next Week’s YC Founder 🎧 Listen on the go: Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP Apple: https://apple.co/4cZ8RMm My other YouTube Channel: https://www.youtube.com/@lobstercap

Factories don’t buy hype. They buy uptime. In this episode, YC-backed Cerrion CEO Karim Saleh explains why industrial computer vision is one of YC’s most underrated AI wedges, and why it can beat flashier “copilot” markets on sheer dollar-weight and defensibility. Karim breaks down how Cerrion deploys camera-based AI agents to detect production issues in real time, why manufacturing is a $20T opportunity, and what most investors get wrong about selling into factories. If you’re building “real-world AI,” this is the playbook… from five paid pilots before code to scaling via workflows, not dashboards. What you’ll learn: Why YC keeps funding industrial computer vision (even across hype cycles)The technical truth: every factory is different… and why VLMs change the gameThe go-to-market wedge that wins: small scope → fast ROI → expansion (NRR)How Cerrion embeds into daily factory workflows (shift meetings, handovers, owners)The biggest investor myth about manufacturing sales cycles, and the real unlockWhy competition-heavy AI markets are a trap (and what to build instead) 📌 Watch now and subscribe for more smart, fast, founder-first interviews. Chapters: 00:00 Introduction to the Manufacturing Market Opportunity 00:38 Welcome to Lobster Talks 01:11 Introducing Karim Saleh and Cerrion 02:16 The Evolution of AI in Manufacturing 03:47 Challenges and Solutions in Industrial AI 05:40 The Impact of US Reindustrialization 08:44 Customer Success and Expansion Strategies 15:44 Fundraising and Market Trends 26:07 Future of Industrial AI and YC Insights 36:07 Conclusion and Final Thoughts 🎧 Listen on the go: Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP Apple: https://apple.co/4cZ8RMm My other YouTube Channel: https://www.youtube.com/@lobstercap

Focus groups used to take 6–8 weeks and cost around $30–50k. Motives is a YC Summer ’25 company that does it in a day, and is doing it the hard way: with real humans, not synthetic personas. In this episode of Lobster Talks, we sit down with Sean (Motives, YC S25) to unpack how AI-native companies are replacing legacy services, why YC is doubling down on agents, and the real tradeoffs of building outside San Francisco. You’ll learn: Why Motives can run focus-group-grade research in 1–2 days vs 6–8 weeksThe truth about synthetic users vs real humans (and where each wins)YC’s real “secret” (spoiler: it’s not a secret) and why it maps perfectly to agentsThe underrated founder problem: sales vs customer success once you have tractionWhy “unlimited research” pricing can create addiction-level usageLondon vs SF: customer density, talent economics, and survivability as a founder 📌 Watch now and subscribe for more smart, fast, founder-first interviews. Chapters: 00:00 Introduction to Motives: Revolutionizing Consumer Research with AI 00:31 Welcome to Lobster Talks: Meet Sean from Motives 01:01 Deep Dive into Motives: How It Works and Its Advantages 01:49 The Evolution of Market Research: From Traditional to AI-Powered 03:52 Comparing Human and AI-Driven Consumer Research 09:58 The YC Experience: Building AI-Native Businesses 14:57 Challenges and Strategies for AI Startups 24:07 Sales and Customer Success in AI Agent Businesses 28:10 Reflecting on Client Payments and Capital 28:35 Testing the Unlimited Plan 28:57 Challenges and Learnings from AI Research 32:14 Quality Control and Automation 34:36 The Importance of Customer Feedback 36:20 Choosing London Over San Francisco 38:16 Advantages of Being in London 44:02 Balancing Work and Personal Life 47:38 Cultural Differences in Tech 52:40 Final Thoughts and Reflections 🎧 Listen on the go: Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP Apple: https://apple.co/4cZ8RMm My other YouTube Channel: https://www.youtube.com/@lobstercap

Everyone’s chasing shiny AI agents. Tanner Jones is quietly using AI to tear out 25% of a state’s rule book, and got Virginia to require his product by law. This episode goes inside Vulcan Technology, a YC startup using AI to map every law and regulation in America, undercut Deloitte and McKinsey, and turn a $4T regulatory burden into a trillion-dollar software market. We talk about Virginia’s $24K cheaper homes, destroying the Big Four with code, how to actually sell into government, and why most “AI for gov” plays are doomed from day one. You’ll learn: How Vulcan uses AI to analyze entire regulatory codes and show what can’t change, what can change, and what must change.The Virginia case study: how building code streamlining translated into ~$24,000 less per new home without touching safety.Why both red and blue states are leaning into AI for regulation, and how YC founders are navigating the politics.What investors get wrong about govtech, ARR, and how to value AI consultancies selling into the state.How Tanner sold to a state government 10 days after incorporation and then raised one of the biggest seeds of his YC batch.What’s overhyped in AI infra, what’s inevitable, and how to avoid building a “solution in search of a problem” in govtech. 📌 Watch now and subscribe for more smart, fast, founder-first interviews. Chapters: 00:00 Introduction: The Challenges of Government Contracts 00:51 Sponsorship Message from Stable 01:51 Introducing Tanner Jones and Vulcan Technology 02:33 The Vision: AI in Government Consulting 06:20 Case Study: Virginia's Regulatory Overhaul 12:26 Bipartisan Appeal of AI Solutions 17:34 Behind the Scenes at Y Combinator 22:28 Navigating Government Sales and Investor Relations 24:51 Scientific Innovation and Emergence of New Companies 25:15 Palantir and Anduril's Impact on Intelligence Agencies 25:37 Targeting Regulatory Agencies and Market Size 26:10 Challenges and Opportunities in Government Contracts 27:37 Passion for Government Reform 28:25 Investor Interest in Govtech and AI 29:30 Challenges for Govtech Startups 32:26 Understanding Government Incentives 34:06 Metrics for Success in Govtech 42:31 Future of AI in Government 46:09 Final Thoughts and Contact Information 🎧 Listen on the go: Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP Apple: https://apple.co/4cZ8RMm My other YouTube Channel: https://www.youtube.com/@lobstercap This podcast episode is sponsored by Stable. Click here to sign up for a virtual address: https://dashboard.usestable.com/onboard/begin?utm_source=youtube&utm_medium=description&utm_campaign=influencer_lobster&promoCode=lobster50