Casey Chu (@caseychu9) / X

Casey Chu

265 posts

Casey Chu

@caseychu9

Researcher at @openai

San Francisco, CA

Joined August 2017

Pinned
Casey Chu
@caseychu9
Jul 17, 2025
We launched ChatGPT Agent today! When tested on a variety of REAL work tasks (expert tasks that might take >10h), we found that its output was human-quality almost 50% of the time Agent puts o3's intelligence into practice - try your work tasks and let us know how it goes!
00:46
OpenAI
@OpenAI
Jul 17, 2025
ChatGPT can now do work for you using its own computer. Introducing ChatGPT agent—a unified agentic system combining Operator’s action-taking remote browser, deep research’s web synthesis, and ChatGPT’s conversational strengths.
16K
Casey Chu
@caseychu9
Nov 20, 2023
OpenAI is nothing without its people :(
57K
Casey Chu
@caseychu9
Nov 21, 2023
1) Vision: when I joined in 2020 I was impressed to hear a coherent plan (from @sama @ilyasut @gdb + others) to build AGI, one that's roughly panning out. leads to everyone pulling in the same direction. at other labs, it felt like people pull randomly
Peter Yang
@petergyang
Nov 21, 2023
The most interesting thing about OpenAI is: How did the company build a culture that not only shipped GPT4, DALLE, and ChatGPT — but also made every staff willing to follow Sam out the door? Do any OpenAI staff want to share?
298K
Casey Chu
@caseychu9
Nov 20, 2023
thanks for posting this - hoping to just pick up where we left off on friday ❤️
Ilya Sutskever
@ilyasut
Nov 20, 2023
I deeply regret my participation in the board's actions. I never intended to harm OpenAI. I love everything we've built together and I will do everything I can to reunite the company.
47K
Casey Chu
@caseychu9
Jan 23, 2025
We launched a research preview of Operator today! It's a model built on top of GPT-4o that can control a browser — it is very early and will make mistakes, but it's a taste of things to come openai.com/index/introduc…
00:00
28K
Casey Chu
@caseychu9
Nov 21, 2023
Replying to @caseychu9
3) No egos: everyone wants to pick up a shovel. do unglamorous work. choose the simple approach, don't bother with complexity unless necessary no matter how pretty the idea is. trust the empirical - if you notice that the model gets better as you scale, why not just keep scaling?
11K
Casey Chu
@caseychu9
Nov 21, 2023
Replying to @caseychu9
2) The whole company is a team: one part incubates, one scales up, one deploys. by the time the public sees, the whole company has contributed. DALL-E was a big chunk of the company. GPT-4 was a whole-company effort. the current crisis? just another whole-company effort.
14K
Casey Chu
@caseychu9
Apr 30, 2020
GAN techniques like Lipschitz constraints + gradient penalties can be seen as adding smoothness to the loss function that the GAN minimizes - includes some neat ideas from convex analysis! Work w/ @ktrmnm @KFukumizu - check out our #ICLR2020 poster at youtu.be/5ZRL06xLeVQ
Casey Chu
@caseychu9
Jul 17, 2025
working on bringing that pass@16 number down to pass@1 💪
Epoch AI
@EpochAIResearch
Jul 17, 2025
Replying to @EpochAIResearch and @OpenAI
We also found that, when allowed 16 tries per problem, ChatGPT agent’s score grew from 27% to 49% on the tier 1-3 set. This suggests that better prompting or scaffolding might result in better performance from current models.
2.8K
Casey Chu
@caseychu9
Jan 31, 2019
Replying to @gabrielpeyre
Check out an interactive version of that diagram that I made! Click and drag on this page to visualize the geodesic between two normal distributions: caseychu.github.io/normal-distrib…
Casey Chu
@caseychu9
Apr 6, 2022
Started reading this chapter on Causality by @victorveitch @alexdamour in @sirbayes's book and love it so far - it's so clearly written! Question about "the ideal gas law doesn't trivially map to SCMs": would it work to define an SCM w/ variables for 2 times, P1/V1/T1/P2/V2/T2?
Casey Chu
@caseychu9
Apr 7, 2022
An amazing experience to work with this team and train stacks and stacks of diffusion models! So exciting to see what this technology can do :) Check out DALL-E 2 if you haven't already! It still blows my mind this is possible 🤯 openai.com/dall-e-2
Prafulla Dhariwal
@prafdhar
Apr 7, 2022
For those who wanna read our paper, check it out here! cdn.openai.com/papers/dall-e-… Huge shout out to the awesome @model_mechanic @unixpickle @caseychu9 @markchen90 and so many more people across teams at OpenAI who made this possible!
Casey Chu
@caseychu9
Jun 6, 2025
Join us in making the next generation of agents both capable and safe! We think that agents will be a big part of how we interact with AI in the future, making it critical that we think carefully about how we build them.
fouad
@fouadmatin
Jun 5, 2025
We're hiring for a new team @OpenAI: Agent Robustness and Control Our goal is to make sure our agents safe and secure during training and deployment. Want to work on some of the hardest problems in AI today? Apply via link in reply or DM me!
3.5K
Casey Chu
@caseychu9
Jul 17, 2025
Replying to @caseychu9
Huge thanks to @gracejkim9 Elizabeth Proehl @michelelwang @marwan_aljubeh @rachelds__ @tejalpatwardhan for putting this eval together, letting us measure Agent's capabilities in realistic work settings 💪
748