No PriorsNo Priors Ep 66 | With Y Combinator President and CEO Garry Tan
At a glance
WHAT IT’S REALLY ABOUT
Garry Tan on YC, AI startup surge, and rebuilding San Francisco
- Garry Tan, President and CEO of Y Combinator, discusses YC’s evolution from a scrappy subculture to a defining institution for global startups, while arguing that its core philosophy—backing highly technical, clear-thinking founders—has stayed constant.
- He explains how AI is reshaping YC’s batches, with about 70% of companies now AI-related, and describes why he thinks “ChatGPT wrapper” is a misleading critique, likening LLMs to core primitives like databases or cloud, not mere features.
- Tan outlines YC’s emphasis on first-hand customer insight, rapid shipping, and social pressure (like group office hours) as drivers of extreme early-stage growth, especially in the current AI wave where many startups achieve rapid revenue traction.
- He also talks about his civic work in San Francisco, arguing for effective governance, strong public education (especially math), and more open, nuanced civic dialogue as essential to sustaining the city’s role as a global tech hub.
IDEAS WORTH REMEMBERING
5 ideasHighly technical founders who communicate clearly are YC’s enduring bet.
Regardless of hype cycles, YC prioritizes deeply technical people who can explain what they’re doing simply and who talk directly to customers, not just investors or other founders.
First-hand customer insight beats second- and third-hand ecosystem chatter.
Tan stresses that talking to users is “first-hand,” other founders are “second-hand,” and investors or media are “third- or fourth-hand,” so the best ideas come from direct engagement with a narrow, specific user segment.
AI ‘wrappers’ can be real businesses, like SaaS atop databases or cloud.
He rejects the idea that using an LLM API makes a startup trivial, comparing it to calling SaaS companies mere ‘MySQL wrappers’—the value is in workflow design, domain integration, and solving specific, painful problems.
Breaking work into precise, automatable steps is key to useful AI products.
Examples like Casetext show that mapping detailed “time and motion” of expert work (e.g., how lawyers read and annotate briefs) into prompts, workflows, and evaluation loops turns general models into real, billable tools.
AI is accelerating early revenue growth for young startups.
Recent YC batches saw average ARR jump from about $6M to over $30M in three months, which Tan attributes to AI’s ability to directly replace or augment expensive, repeatable knowledge work that companies already outsource or staff.
WORDS WORTH SAVING
5 quotesYC started more like a punk club than an institution.
— Garry Tan
Talking to customers is first-hand; talking to investors is third-hand.
— Garry Tan
Calling AI products ‘ChatGPT wrappers’ is like calling SaaS ‘MySQL wrappers.’
— Garry Tan
If you can run fast for a long time, compounding takes care of the rest.
— Garry Tan
We need to save ambitious, optimistic people from investment banking and middle management.
— Garry Tan
High quality AI-generated summary created from speaker-labeled transcript.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome