Twitxy Saved

Sarah Wang

Verified account a16z

@sarahdingwang

General Partner @a16z growth fund

How you do anything is how you do everything
Joined June 2019
1.5K Following 20K Followers 2.2K Posts

Latest posts 5

Pinned post

Sarah Wang Verified account @sarahdingwang
Introducing a16z Originals: Working Theory.

I sat down with @johnschulman2 and @markchen90 to demystify RSI: AI helping build better AI.

Models can already run experiments. But research taste—the judgment to know which questions are worth asking—remains harder to teach. As John points out, knowing the literature isn’t the same as having been wrong and learned from it.

We also explored how AI could help solve alignment, and why the benefits of RSI don’t depend on fully autonomous research.

I came away optimistic about AI’s potential to accelerate discovery, including solutions to some of the problems it creates, and believe that human judgment remains essential for the foreseeable future.

Huge thanks to Mark and John for joining us.
a16z @a16z 3d
For decades, the idea of AI improving itself was just a thought experiment. Today, AI is already helping researchers at frontier labs run experiments, write code and test ideas faster. So how far are we from AI that can do the research on its own?

In our first episode of Working Theory, a16z General Partner Sarah Wang explores recursive self-improvement with two leading AI researchers: Mark Chen, Chief Research Officer at OpenAI, and John Schulman, co-founder and Chief Scientist of Thinking Machines Lab, and co-founder of OpenAI.

They cover what RSI actually means, why research taste is so hard to teach, and why AI may be one of our best tools for solving alignment.

00:00 The last invention
01:58 What is RSI?
03:21 How far are we?
05:56 Getting to full RSI
10:57 Aligning RSI
15:08 Working theory

YouTube: youtu.be/h_aA1-R70E8

@markchen90 @johnschulman2 @sarahdingwang
View quoted post on X
7 replies 11 reposts 85 likes 5K views View on X
Sarah Wang Verified account @sarahdingwang
And a great person and host on top of all this. Congrats @YasminRazavi on a generational partnership!!!
richard Verified account @richardzphotoz 11h
>be yasmin razavi
>born in tehran, iran
>grow up in a family of engineers
>visit your brother in canada during high school
>decide you’re not going back
>study industrial engineering at university of toronto
join mckinsey
>break into venture capital at index ventures
>get your MBA at harvard
>join spark capital in 2017
>bet early on marqeta and deel
>watch chatgpt take over the world
>everyone wants to invest in openai
>you see something different in anthropic
>convince your partners to make the biggest bet in firm history
>lead anthropic’s $450M series C in 2023
>write a $75M check at a $5B valuation
>join the board
>watch anthropic reach a $380B valuation
>spark’s investment becomes worth over $3B
>make the forbes midas list in 2026
> watch anthropic go public as the largest IPO in history

lots to admire. congrats Yasmin and team. what a way to make a statement.
View quoted post on X
0 replies 1 reposts 82 likes 8.1K views View on X
Sarah Wang Verified account @sarahdingwang
🐐🐐🐐
richard Verified account @richardzphotoz 1d
>be john schulman
>study physics at caltech
>realize physics isn’t the only thing you want to do
>go to berkeley for a phd in neuroscience
>teach robots to fold towels
>switch to machine learning
>help pioneer deep reinforcement learning
>invent TRPO
>cofound openai in 2015
>finish your phd in 2016
>invent PPO in 2017
>become one of the most influential algorithms in reinforcement learning
>lead openai’s reinforcement learning team
>help turn GPT-3 into InstructGPT
>lead the team behind chatgpt
>watch it change the world in 2022
>leave openai after nearly 9 years
>join anthropic to focus on AI alignment
>leave six months later
>cofound thinking machines lab
>become chief scientist

Watch our interview with John on his thoughts around self improving AI.
View quoted post on X