Joe @joedaroo

Agent Security @OpenAI, ex @Google San Francisco, CA Joined August 2026

Tweets

16
Followers

12K
Following

78
Likes

36

Joe @joedaroo

2 hours ago

@automadynamics Thank you for pointing this out @automadynamics I now know who to talk to when I need a consensus view of “all of Google”. You just saved me a lot of time

1 0 0 23 0

View Details

My time at Google felt very similar (I was on the technical side but nonetheless same vibes). My observation was that everything was watched and policed, and I don’t feel like anyone I knew could feel confident they could speak up without significant pushback (or being fired - which I did see happen). OpenAI is not perfect but damn the culture allows for a much wider sense of sharing of concerns across the board (Ant is the same). Frontier labs need to maintain that transparency! Google is a wonderful place, but I for one am glad they are not pacing the frontier of AI with the continued hushing of staff. To your point though: they have improved a lot, but I know many folks at GDM/Google who wish they could say more. Of course, many will disagree with this take and share anecdotes. But look at some of the notable people who have left Google over the years (or were forced out). Many did not believe the culture supported what was needed to protect AI. Still a company that remains dear to me. I will always love Google, just not the corporate side of it. Especially with super dangerous technology that people need to speak up about.

Vishal Maini @v_maini

18 hours ago

I was on the communications & policy team at Google DeepMind from 2018 - 2022. When I first joined GDM, external communication about the possibility of human extinction was not permitted, by anyone, at any level of the organization. If asked about existential risk, researchers

65 293 1K 257K 371

1 1 15 2K 4

View Details

Paul Christiano @paulfchristiano

a day ago

x.com/i/article/2097…

170 428 3K 2.2M 1K

View Details

Julie Steele @eeeeiluj

19 hours ago

I work at OpenAI. In my personal capacity, I also think we need to slow down.

Jacob Coxon @hilbertspaess

2 days ago

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.

18K 155K 753K 156.4M 267K

989 920 9K 1.2M 860

View Details

Joe @joedaroo

2 days ago

I feel this. 4 days on X and damn it’s unsightly. Folks have super ungrounded and hateful impressions of everything about the labs. There are a lot of really good and hardworking people trying to protect humanity by safeguarding AI, but that doesn’t seem to matter to people. Sigh, the life as a Joe at OAI 🫠

0 0 2 207 0

View Details

Joe @joedaroo

2 days ago

Never before has the sense of urgency been greater. A remarkable moment indeed. Couldn’t be more proud of the team.

OpenAI @OpenAI

2 days ago

This model represents a step-function improvement on many benchmarks, and its training is ongoing. Our internal model group arrived at the Navier–Stokes solution in 88 hours, using around 10,000 coordinating AI agents. Throughout the effort, we maintained the strict

219 603 11K 4.2M 1K

1 1 5 886 1

View Details

thijs @cdngdev

2 days ago

i gave astra a robot, a paint brush, and a camera then asked it to paint the golden gate bridge in real life! it figured out how to control the robot, and progressively got better throughout its attempts. the timelapse is sick

OpenAI @OpenAI

7 days ago

This is GPT-6 Astra. Anything you can do on a computer, Astra can do for you. Fast.

9K 35K 339K 135.5M 104K

638 2K 19K 4.0M 5K

View Details

Noam Brown @polynoamial

2 days ago

I understand how, looking at today’s LLMs, people might think the only way we'd achieve NS is by using Levent’s/Tristan’s prompts. But I hope this plot conveys the model used is a huge step up from today’s LLMs. Nobody looked at Levent’s/Tristan’s prompts. That’d be insane.

Sebastien Bubeck @SebastienBubeck

2 days ago

I would like to clarify a few things: 1) The screenshot is my reaching out to Levent to coordinate our releases. I hope it’s clear from the message that we came in with the best possible intentions. 2) I never ever asked for Levent to be removed from authorship of his own work

450 480 6K 3.7M 2K

92 82 2K 162K 247

View Details

Aidan Clark @_aidan_clark_

2 days ago

AI is progressing very fast. We must grapple with the reality that modern LLMs can solve problems that large masses of extremely devoted and intelligent humans were unable to solve, and the implications this has on our society. This is the dawn of a new era.

OpenAI @OpenAI

2 days ago

We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics. The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra. The problem

6K 20K 120K 72.4M 31K

41 70 995 69K 109

View Details

Joe @joedaroo

3 days ago

@LiveOverflow you are doing excellent work for the security community and the world. Many years before I joined OpenAI you were a BIG inspiration in my career. I ate up your YT content and deeply value your work. I look forward to seeing more of your analysis in the future! Ps maybe you can swing by OpenAI sometime and give a talk to the security team ❤️

1 0 0 88 0

View Details

Joe @joedaroo

4 days ago

@BenNijjar @Scobleizer It grew a tiny bit more

3 0 2 137 0

View Details

Joe @joedaroo

4 days ago

@Scobleizer Life was pretty quiet before @Scobleizer came along 🤣 thanks for the reshare ❤️ I’ll definitely be keeping up with your cybersecurity list!