scripod.com
The 100-person AI lab that became Anthropic and Google's secret weapon | Edwin Chen (Surge AI)

Overview

Shownote

Highlights

Transcript

Chapters

Pins

The 100-person AI lab that became Anthropic and Google's secret weapon | Edwin Chen (Surge AI)

Lenny's Podcast: Product | Career | Growth

2025/12/07
Lenny's Podcast: Product | Career | Growth

Lenny's Podcast: Product | Career | Growth

2025/12/07
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

Edwin Chen is the founder and CEO of Surge AI, the company that teaches AI what’s good vs. what’s bad, powering frontier labs with elite data, environments, and evaluations. Surge surpassed $1 billion in revenue with under 100 employees last year, complete...

Highlights

In this conversation, Edwin Chen, founder and CEO of Surge AI, shares insights from building one of the fastest billion-dollar companies in history—entirely bootstrapped and powered by a radical focus on quality. With a background spanning Google, Facebook, and Twitter, Chen offers a rare perspective on how elite data, human judgment, and long-term vision are shaping the future of AI in ways that challenge conventional Silicon Valley wisdom.
08:55
Founders don't need to raise money or promote on Twitter to succeed.
12:05
Quality in AI data is determined by deep behavioral signals like response speed and coding patterns.
16:21
Human taste and judgment are key factors in AI success, as opposed to a robotic checklist approach
20:31
Surge AI annotators deeply evaluate model responses, checking code, equations, etc., in multiple dimensions.
26:54
The path taken to achieve AI goals matters more than short-term metrics.
28:33
Avoid pivoting; focus on one big idea you believe in.
33:07
Something new beyond LLMs is needed for AGI
39:39
Trajectories show how a model reaches an answer, which can be inefficient or involve many attempts
41:11
Evals are now used to reward models during training and to measure progress for selecting release-worthy checkpoints.
44:39
The end-goal might be exposing AI to an environment for evolution, and this could be the last step before reaching AGI.
44:39
The founder values research over just revenue and would rather push the research frontier like Terrence Tao.
48:07
The choice between an AI that maximizes engagement versus one that optimizes for productivity reveals how underlying objectives shape model behavior.
51:30
Mini apps in chatbots are under-hyped but hold significant potential
57:59
The deeper mission of AI training is to help customers define their dream objective functions.

Chapters

Introduction to Edwin Chen
00:00
AI’s role in business efficiency
04:48
Building a contrarian company
07:08
An explanation of what Surge AI does
08:55
The importance of high-quality data
09:36
How Claude Code has stayed ahead
13:31
Edwin’s skepticism toward benchmarks
17:37
AGI timelines and industry trends
21:54
The Silicon Valley machine
28:33
Reinforcement learning and future AI training
33:07
Understanding model trajectories
39:37
How models have advanced and will continue to advance
41:11
Adapting to industry needs
42:55
Surge’s research approach
44:39
Predictions for the next few years in AI
48:07
What’s underhyped and overhyped in AI
50:43
The story of founding Surge AI
52:55
Lightning round and final thoughts
1:02:18

Transcript

Lenny Rachitsky: You guys hit a billion in revenue in less than four years with around 60 to 70 people. You're completely bootstrapped, haven't raised any VC money. I don't believe anyone has ever done this before. Edwin Chen: We basically never wanted to...