40:36Ask These Questions Before Starting An AI Startup
From Y Combinator · Published Oct 23, 2025 · Watch on YouTube
TL;DR
The speaker, an Anthropic alignment research lead and former YC founder, expresses deep uncertainty about AI's trajectory and argues founders should plan for AGI within 2–3 years. He presents a series of open questions about product strategy (retrofit vs. AI‑native, on‑demand code, trust), team culture, alignment economics, data moats, and the need for AI‑powered auditing to rebuild trust.
Key insights
- Founders should plan two years ahead assuming AGI will arrive, not just six months.
- The buy side (enterprises) will also be armed with AI agents, accelerating adoption and potentially bypassing SaaS products entirely by building custom software in‑house with tools like Claude Code.
- On‑demand code generation could eliminate the app model: users prompt the system to build a bespoke feature at runtime, but this requires high trust in model reliability and security controls.
Want the full analysis - every claim cited to the second it was said?
This page only shows a teaser. Sign up to chat with the complete, cited breakdown of "Ask These Questions Before Starting An AI Startup".