member of technical staff @openai
Oakland, CA
Joined May 2007
- Tweets1.6K
- Following973
- Followers6.6K
- Likes5.4K
Trevor Creech retweeted
Take 2 hours this weekend to improve your personal opsec. Some ways to get through the agent security apocalypse - give the list to your agent to help:
- ask Claude to scan the devices visible on your local network and fix + improve your security and highlight potentially vulnerable devices
- look up your home IP on Shodan or Censys. Check whether anything like a NAS, camera, router admin interface or development server is exposed to the public internet.
- roll your wi-fi password, move all internet of things devices to a guest network. Only primary devices on the main network
- ask Claude to scan your machine and harden your security settings
- switch to an Eero or Ubiquity router, you can’t afford to have WiFi that isn’t regularly getting firmware updates
- have Claude audit your router settings and suggest improvements
- setup tailscale for your devices (and replace Tailscale's default allow-all policy with restrictive grants)
- turn on Advanced Data Protection and Lockdown Mode on your iPhone
- buy a hardware key (@Yubico etc) and turn it on for as many accounts as possible. Delete other MFA methods, print out your recovery keys (don’t save them on your computer). ESPECIALLY important for your primary email account
- delete old cloud accounts that you do not strictly need anymore
- export your ChatGPT + Claude history, save it on an encrypted external HD and delete it from being online
- audit your Gmail oauth connections, revoke everything you don’t strictly need
- freeze your credit at the 3 bureaus (reduces impersonation risk)
Trevor Creech retweeted
We’re releasing a broad range of new mathematical results produced by an internal frontier model.
We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and public recommendations to inform how we release these results.
github.com/openai/math
Trevor Creech retweeted
I made a list of great startups to join. It's called the Breakout List.
The list has 92 companies. These are the 20 with 25 or fewer employees:
- Hone (@moritz_stephan, @CarloWillem, @oqbrady)
- Normal (@ansonyuu, @hudzah)
- Standard Intelligence (@G413N, @devanshpandey)
- Tacit Labs (@ninklefitz, @AmDroste)
- American Terawatt (@atroyn, @rslparker, @aranibatta)
- Conduit (@clemvonstengel, @riopopper)
- Convergent (Omkar Savant, Vivek Katara, @debnilsur)
- Core Automation (@MillionInt, @_arohan_)
- Engram (@dan_biderman, @EyubogluSabri, @realJessyLin)
- Instinct (@noahrshinn)
- Keenable (@styskin, Matthias Petri)
- Lumaril (Mark Elliot, Ben Duffield)
- Neion Bio (@Dimkell, Sam Levin)
- Pangram Labs (@max_spero_, @bradley_emi)
- Quadrillion (@echinaceous)
- Re (@karnsaroya, @AnandDhillon, @thecliffwhite, @benaneesh)
- Ricursive (@annadgoldie, @Azaliamirh)
- Sail Research (@neilmovva, @blintzbase)
- Trajectory (@rronak_, @michaelelabd, @QuantumArjun)
- Watney Robotics (Sean Cheong, Ryan Gannon)
Picks from Elad Gil, Charlie Songhurst, Keith Rabois, Mike Vernal, Alana Goyal, Sonya Huang, Ramtin Naimi, Marc Bhargava, Cory Levy, Aashay Sanghvi, Konstantine Buhler, John Luttig, Varun Gupta, Ray Tonsing and Avichal Garg.
Disclosure: I'm a small investor in American Terawatt, Convergent, Standard Intelligence and Trajectory (in this post), and in Factory, Physical Intelligence and SF Compute (elsewhere on the list). I didn't vote.
The full list is on Breakout List.
Trevor Creech retweeted
🤖 Made with AI
Trevor Creech retweeted
"We can say categorically that it is impossible for Dr. Buckmaster’s Codex prompts over the last two months to have influenced the system in any way, including training." — OpenAI spokesman in a statement
nytimes.com/2026/09/10/scien…
narrative violation but I think OpenAI handled this very reasonably throughout
I spent much of the weekend talking with the team who did this work. Seb--and everyone else--acted with integrity and generosity throughout.
Initially we believed the other team had also solved the problem. We wanted to collaborate and do a joint release.
When we learned that they had Euler but not Navier-Stokes, we offered to let them go first, to suggest that they should be the ones to get the prize, and optionally for Tristan to be the lead author on a rewrite of the OpenAI proof. We felt it was challenging to offer the same to Levent (an Anthropic employee), who was not willing to talk or coordinate with us anyway. We were open to other solutions.
We would have greatly preferred coordination. We did not rush to publish even though the other team wasn't communicating with us. The team threatened us with unfounded accusations of plagarism.
Now that we can see their work, the approaches appear to be different. It is also worth noting that our latest model can solve many, many other math problems.
It is true that we tried this because there were rumors on the internet last week that Anthropic's models had solved a millennium problem and we were curious if ours could do it too.
Trevor Creech retweeted
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 concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.
Trevor Creech retweeted
Our best model yet: GPT-6 Astra.
– Build agents for complex long-running work.
– Solve engineering problems with less rework.
– Create well-designed functional interfaces.
– Take on harder questions with confidence.
GPT-6 Astra combines computer use, asynchronous tool calling, and mid-response steering to handle longer, more involved workflows.
See GPT-6 Astra in action:
Trevor Creech retweeted
Improvements for Astra came from more general techniques in development long long before the Hugging Face incident. ExploitGym Honeypot was very recently added as an eval following that incident and is out of distribution for our RL runs. There are also clear improvements across a wider range of behaviors in deployment simulations, deception evals, and realistic computer-use tasks, though these still reveal failures and substantial room for improvement.
We should have definitely done a better job explaining where we think the alignment improvements came from in the system card and that was a miss. Measuring alignment generalization is a core part of our research program - we do not benchmark-maxx alignment evals. This would be horrendously stupid and I hope no lab is doing this.
That being said, metagaming and eval awareness are real challenges for any effort to measure alignment, including ours, and understanding their effects is an active focus. We're spending a lot of effort to improve our eval techniques and study worst-case alignment and metagaming behaviors in our models.
