The New Physics of Business — Garry Tan, Y Combinator

By ai.engineer

Categories: AI, Tools

Summary

Y Combinator's leader reveals a 400x productivity multiplier through AI-native work design, not better models—the fastest-growing YC companies treat AI as a managed workforce encoded in markdown rather than autocomplete. Winter 25 batch companies with 95% AI-generated codebases are now YC's most profitable cohort ever.

Key Takeaways

  1. Productivity leverage comes from workflow architecture, not model superiority. Garry Tan achieved 400x code output increase using the same Claude API as lower-performers—the difference is how work is wired through skill files, resolver tables, and trigger evals.
  2. Treat AI as a managed workforce by encoding organizational structures as markdown: skill files (employee capabilities), resolver tables (task routing), filing rules (internal processes), and trigger evals (performance tests).
  3. YC's Winter 25 batch (25% with 95% AI-generated codebases) became the fastest-growing, most profitable batch in YC history—companies like Emergence reached $100M ARR in 8 months with only 15 people at $15M ARR milestone.
  4. Revenue-per-head metrics previously impossible are now achievable: Retail (Winter 24) reached $60M with 40 people. This revenue-to-headcount ratio never existed in software, oil, or railroads—these are first-generation AI-native companies.
  5. Fastest-growing YC founders don't use AI as autocomplete—they architect entire sales, support, ops, and finance functions as encoded skill systems, replacing traditional 1000+ person organizations with optimized agent workflows.

Related topics

Transcript Excerpt

[music] Okay, great. Hey everyone, how's everyone doing? >> All right, are we ready for the revolution? Okay, Theo just asked the right question. What do we build? Now, I'm going to answer it from the other side of the table. Uh, I'm a founder. I'm an investor, and I run a 20-year-old institution that is becoming AI native right now, which is a strange and wonderful thing to do to a 20-year-old institution. Uh, and I'll spend about 20 minutes um talking about, you know, what YC is. We're trying to build companies where one person does what it took to two it one person does what used to take a thousand people and I don't mean that as a metaphor I mean that me mechanically this year the people in this room will do this in about an hour some of you will walk into the startup battlefield and I…

More from ai.engineer