Building at @OpenAI. Previously software engineer @Statsig and @Microsoft. Statistical learning, ML systems. Based in Bellevue.
Joined January 2015
- Tweets370
- Following2
- Followers12
- Likes4
Control Room is just one part of our holistic approach for safe frontier model training and deployment.
We built Control Room, a central system to track model workloads across our research infrastructure and stop them quickly when safety or security concerns arise.
It connects every research inference workload, including agentic sampling for model training and evaluation, data processing, and more. This works together with our misalignment monitoring systems, and halts model output within 35 seconds of a decision to intervene to stop or pause a job.
Control Room is just one part of our holistic approach for safe frontier model training and deployment.
openai.com/index/control-roo…
1
1
7
absolutely incredible
GPT-6.2 Sol is already showing up for some ChatGPT users, and just look at what it's creating.
These aren't screenshots from a game. They're Three.js scenes generated from a single prompt, with everything packed into ONE HTML file.
I've been waiting to see what Sol could do. And now I think we're finally going to see a worthy competitor to Opus 5.5 for Plus users.
12
Wild stat. We are now generate more than 4B images per week across ChatGPT and the API, up from the more than 3B per week only 8 weeks ago.
Images, upgraded—introducing ChatGPT Images 2.6.
Now with GIFs and reference images.
Plus, two new models are coming to the API with 50% faster image generation.
1
1
9
And yet, and this makes me sad, people are so against AI. We have so much work to do friends
OpenAI's Greg Brockman says AI could make death a choice and predicts significant progress toward curing diseases within the next 12 to 18 months.
Once science fiction, today reality. And yet, people oppose AI.
4
These indi AI will aggregate.
Sam Altman just admitted why the smartest people at OpenAI keep walking out to build their own AI.
Ilya. Dario. Elon. All of them were there early, and all of them left to start rival labs.
His explanation is brutal:
"I think there is something that happens in many people's brains where they only trust themselves to build this because the stakes feel so high."
And this is how he sums up what they're thinking:
"His approach is bad and so I can't let power get concentrated in his hands."
That's the real race now. Not for the best model, but for who gets to hold the most powerful technology ever built.
OpenAI → Anthropic → SSI → xAI → Everyone Trusting Only Themselves
2 minutes that explain the whole AI race better than any headline.
Watch till the end, where he says what building this has done to his own life.
1
1
15
I can confidently say that humans were not very good at writing secure infrastructure. The endstate will be hypersecure infrastructure and software for all, but we're going to go through some turbulence before we reach that.
7
That said, we should note that this sort of headline does not distinguish how severe each vulnerability is ... LLMs can find sooooo many vulnerabilities in any complex real-world codebase, and one of the additional AI problems we face now is to use AI to prioritize which of these are most severe and most urgent to fix... Will be good for OpenAI to address this problem next...
Our security researchers built an autonomous cyber vulnerability discovery, exploitation, and patching pipeline.
In a single run with an upcoming cyber model, it validated and generated patches for 948 local privilege escalation vulnerabilities in the Linux kernel at a total API cost of $1.2 million.
4
People should always be in control of the future.
People should always be in control of the future.
OpenAI is against the notion that it would be acceptable for people to be relegated to a role of doing things that “AI systems think makes sense to reward in humans”.
People should not cede control of the future to superintelligence, and that requirement should ultimately constrain AI progress.
2
LifeVal👀
For AGI to benefit all of humanity, it should help people with the everyday tasks that matter most to them. So we built LifeVal, a new benchmark shaped by feedback from thousands of ChatGPT users on where they most want help from a personal assistant—from finding childcare to accessing government services to managing personal finances.
Over the past year, we’ve used LifeVal to measure and improve how well our models handle these everyday tasks.
openai.com/index/lifeval/
2
a new Astra is coming soon.
OpenAI is quietly A/B testing a new variant called GPT-6.2 Astra, and it's already showing up in Codex usage analytics.
The model picker still says GPT-6 Astra, but the usage breakdown for some accounts now lists gpt-6.2-astra separately. Multiple users are seeing the exact same thing, so this isn't a glitch. It's a gray rollout, and a new Astra is coming soon.
3
J'utilise chatgpt tous les jours depuis plus d'un an... J'avais jamais envisagé d'aller ailleurs car la qualité d'écriture et ce qu'on pouvait faire avec était juste génial.
Bad news for ChatGPT users: starting in December, Plus, Pro and Business limits will be shared across ChatGPT and Codex.
OpenAI quietly updated its help page. Usage will now be measured in tokens, so bigger models and harder tasks burn through your allowance faster. Only Instant chat stays free of the limit. Hit the cap and you wait, upgrade or buy credits. Heavy Codex users may feel this one.
17
that humanity deserves fundamental knowledge.
Sam Altman admits OpenAI let its AI-generated math research get "out of control for one week." He says the company released hundreds of papers that overwhelmed mathematicians.
However, Altman refuses to stop, arguing that humanity deserves fundamental knowledge. He adds that OpenAI has already spent hundreds of millions of dollars on this research.
22
GPT-6.2 Bel is releasing this week
GPT-6.2 Bel is releasing this week, on the final day of Tibos 28-day shipping spree.
The model ID gpt-6.2-betelgeuse was spotted in an internal tracker. A verified OpenAI insider claims it scores 72 on Artificial Analysis, 13 points above Carbon.
Reportedly 12 trillion parameters with $20 input and $100 output per million tokens.
And thats just one agent. OpenAI used 10,000 agents for its Navier-Stokes research.
57