BUILDING CAPABILITY
AI Builders
Start with a real problem, learn to make AI carry complex work reliably, and turn the result into projects, workflows, and systems worth keeping.
Courses · projects · hands-on practice
Explore AI Builders
Yuzheng Sun/课代表立正
From economics to AI: can a judgment change the result?
I’m Yuzheng Sun, also known as 课代表立正. I have a PhD in Economics from Cornell and have worked as an economist, data scientist, and technology leader. Today I’m based in Seattle, where I teach AI, build communities, write books, and host long-form conversations.
In conversation with Ben Gilbert and David Rosenthal of Acquired · Significance Summit
200+
conversations with researchers, founders, and operators
3,000+
paying learners across AI, data, and growth
WSJ 2025
featured in a CIO Journal reading list
400K+
cross-platform audience
Figures last checked July 2026
THE THROUGHLINE
I changed roles several times. The thing I cared about kept getting more concrete.
Economics trained me to ask about incentives and causality. At Amazon and Meta, a model mattered only if it changed a decision. Leading a 30-person data and AI team at Tencent taught me that one person understanding the problem is not enough—the judgment has to travel through an organization. At Statsig, experimentation became infrastructure for continuous correction.
Courses, communities, books, interviews, and enterprise work may look different, but I still approach them the same way: understand the problem, test the idea in real work, and pass on what survives.
Cornell
PhD, Economics
Amazon
Economist
Meta
Data Scientist
Tencent IEG
Vice Director, Data & AI
Statsig
Principal DS · Sole Evangelist
Superlinear
Founder
A FEW WORKING VIEWS
I would rather state a view clearly—and let reality overturn it.
01A good world model matters more than knowing more facts.The real test is whether you can explain why something happened—and change your model when reality refuses to cooperate.World models02Strong opinions, weakly held: state the view clearly, then welcome a better one.I like being challenged into changing my mind. That works only when assumptions and evidence are on the table and the discussion can move the work forward.Strong opinions, weakly held03In the AI era, the most durable move is from user to builder.As tools get cheaper, the valuable work moves upward: choosing the problem, designing the system, and turning a one-off result into something reusable.Users to builders
CURRENT WORK
I spend most of my time on three kinds of work.
Courses help people build capability. Community gives them people to keep working alongside. Enterprise projects take the same methods into the messiness of a real team.
LONG-TERM ENVIRONMENT
Stay Superlinear
A paid membership for people already using AI and bringing real expertise. Members ask serious questions, sharpen their work, and find people worth building alongside.
Paid membership · co-led with Yage
Explore Stay SuperlinearORGANIZATIONS
Enterprise AI transformation
Talks, workshops, and custom programs for teams including Tencent, DoorDash, Pinterest, 1Password, and Amazon. The work starts with what a team actually needs to change, not a tour of AI tools.
Talks · workshops · custom programs
Explore enterprise work
VR
“Yuzheng distills years of product growth wisdom into actionable insight—helping data scientists surface decisive signals, PMs turn numbers into strategy, and founders find a repeatable path to compounding PMF.”
CONVERSATIONS
I keep looking for people who have actually built the thing.
Across 200+ public conversations, I am less interested in collecting conclusions than in understanding how a judgment formed, what evidence supports it, and where it might fail.
These conversations reach a 400K+ audience across YouTube, Bilibili, and Xiaohongshu, including engineers, data scientists, founders, investors, and operators across China and the U.S.
Yangqing Jia
PyTorch co-creator · Founder, Lepton AI
Tian Yuandong
AI researcher · Former Meta AI / FAIR
Howie Xu
Chief AI Officer, Gen · Stanford GSB guest lecturer
Yang Ying · Tulong Dashihua
Neuroscience PhD · Serial entrepreneur
Shuchao Bi
Head of Post-Training · Multimodal, OpenAI
Reynold Xin
Co-founder, Databricks
Vijaye Raji
Founder, Statsig · CTO of Applications, OpenAI
Ryo Lu
Head of Design, Cursor
Gergely Orosz
Founder, The Pragmatic Engineer
BOOKS
One book asks how products find direction and grow. The other asks how work becomes capability.
Growth Data Analytics Playbook covers product-market fit, metrics, and experimentation. 真本事 is about turning work into capability, leverage, and income of your own.
Explore both books
INDEPENDENT SIGNAL
“Yuzheng has built an AI education community that is carefully curated, practical, and unusually useful for people doing real work.”
“Yuzheng shares more than tools. He teaches a philosophy of thinking—the rarest and most valuable thing in the AI era. Even as a scientist, I found his course deeply illuminating.”
“Yuzheng is a true AI-native pioneer. With his distinctive way of thinking, he helps more people achieve superlinear growth in the AI era.”
FREE COMMUNITY
Many ideas only became useful after people in Superlinear Academy tried them, questioned them, and changed them.
It is the free community I continue to maintain: a place to see what people are building and bring a real question of your own.