The 100-person AI lab that became Anthropic and Google's secret weapon | Edwin Chen (Surge AI)
Source
Lenny Rachitsky
Author
Lenny Rachitsky
Date
Terms in this piece · Glossary
eval — A repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.
benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
Why it matters
Surge AI supplies the human data, evalA repeatable test for AI quality — a set of tasks plus scoring — used the way software teams use test suites, because model output is too variable to judge by eyeballing.Full definition →, and RL environments behind leading models, and this conversation explains why benchmarkA standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.Full definition →-chasing misleads training and why RL environments are becoming the key lever — useful context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition → for anyone building or evaluating AI systems.
Key quotes
“We're basically teaching our models to chase dopamine
instead of truth.”
Edwin Chen
“it's kind of crazy that these models can
win IMO gold medals, but they still
have trouble parsing PDFs.”
Edwin Chen
“It's literally optimizing your models
for the types of people who buy tablets at the grocery store.”
Edwin Chen
“I would rather be Terrence Tao than Warren Buffett”
Edwin Chen
“You can actually build a successful company
by simply building something so good
that it cuts through all that noise.”
Edwin Chen
Chapters
00:00Introduction to Edwin Chen
04:48AI's role in business efficiency
07:08Building a contrarian company
08:55An explanation of what Surge AI does
09:36The importance of high-quality data
13:31How Claude Code has stayed ahead
17:37Edwin's skepticism toward benchmarks
21:54AGI timelines and industry trends
28:33The Silicon Valley machine
33:07Reinforcement learning and future AI training
39:37Understanding model trajectories
41:11How models have advanced and will continue to advance
podcastWhy humans are AI’s biggest bottleneck (and what’s coming in 2026) | Alexander Embiricos (OpenAI Codex Product Lead)
podcastThe coming AI security crisis (and what to do about it) | Sander Schulhoff
podcastWhy most AI products fail: Lessons from 50+ AI deployments at OpenAI, Google, and Amazon
podcastThe non-technical PM’s guide to building with Cursor | Zevi Arnovitz (Meta)
podcast“Engineers are becoming sorcerers” | The future of software development with OpenAI’s Sherwin WuLenny Rachitsky
podcastAI’s third era: the rise of persistent AI coworkers | Tara Seshan (OpenAI’s product lead)Lenny Rachitsky