@LinusMixson i was on a call during the first announcement where people were trying very hard to say the precise lawyerly truth having not looked into where codex (opt in) data was used and whether it made it into training runs. it's now clear it would have been impossible, the chances are 0

Life update: I recently joined @OpenAI! And in my first 2 weeks, I helped launch the new Agents API - we move at crazy fast speed. Instead of building your own agent harness and managing your own agent infrastructure, you can use our Agents API! If you have any feedback, DM or reply to me!

Today we're launching the Agents API, a brand new way to build Agents in the cloud, backed by the Codex harness. Bring along all your favorite tools and connectors, connect it to any sandbox, and let Astra cook. Can't wait to see what you whip up 👨‍🍳

I am pleased to see that Ant’s new Claude Code has a version of background computer use on par with the version from Codex from last May this year. Nice work! Shipping great features first turns out to be a great way to encourage other labs to ship too. We will continue to do this until other labs pay more attention to shipping. This is good for everyone and there is a lot of room left to go! We solved computer use in practice for GPT models about four months ago. But computer use is a significantly awesome thing no matter what model you use, and it is important that the industry similarly spends more effort to train their models to be great at computer use, among other elements of model capabilities. As models become more capable and central to businesses and economies, the value of computer use only increases. We should be taking it seriously, and doing the right thing for our customers and the world.

@ChaseLochmiller @OpenAI GPT-6 Astra, trained on ~100K+ NVIDIA Grace Blackwell NVLink72. From ChatGPT to o1 to Astra in 4 years. AGI has arrived. Congratulations @OpenAI team. 400K GPUs coming online next.