Skip to main content
AI RADARModels
357•08/26/2026, 20:45•2 min read

Interview with Thibaut: Head of Codex/ChatGPT at OpenAI Released

Executive TL;DR30s read

An interview with Thibaut, the head of Codex/ChatGPT at OpenAI, reveals upcoming changes in OpenAI technologies.

Interview with Thibaut: Head of Codex/ChatGPT at OpenAI Released

This is the same insider guy who constantly hints at releases and announces limit resets. You might not recognize him from the interview snippet, but it's him (life at OpenAI has taken its toll, and he looks little like his avatar from X).

He made a powerful statement: in 2-3 months, the current Codex will seem primitive. Agents based on the next generation of OpenAI models (Astra) will stop working locally; they will require a cloud where many more parallel processes will run than an ordinary laptop can handle.

At the same time, the speed of agents will significantly increase, and what we currently call Ultra Fast (750 tokens per second and more) will become the default setting.

Moreover, we are expecting an even tighter integration of Codex and ChatGPT. Future models will work out of the box on the same harness for coding, research, design, searching, etc. Thus, the ultimate goal is a single highly personalized application of Personal AGI, whose interface will change depending on what you are specifically doing.

The full interview lasts 45 minutes; you can watch it here: https://youtu.be/4qjEgPojjzM

Why it matters

Analysis

This interview is significant as it unveils new horizons in the development of OpenAI technologies that could fundamentally change how we work with AI. The announced changes may impact the future of programming and interaction with artificial intelligence.

Discuss in community

Share your questions and insights with developers

Join Telegram
+5 Points
← Back to news feed

Related in Ecosystem

Practical guides, core terminology, and agent skills