Vibeleaderboard
← All Intel
Intel / video

Codex for Everyday Work: AI Agents Beyond Coding

Source
youtube.com
Author
OpenAI
Date

Context

In an OpenAI Forum conversation, Thibault, introduced as head of Codex at OpenAI, says most tasks run in Codex are now not coding. The explanation given is that software engineers spend perhaps 20 to 30 percent of their time actually writing code. The rest goes to tickets, discussions, bug investigation and gathering information, and early adopters were already using Codex for that surrounding work.

Thibault also describes a change in design. The first public version ran in the cloud and opened pull requests, but Thibault says setup was too hard and the models were not reliable enough for long tasks, so the team moved Codex to people's own computers. Giving the access to material such as documents and team tools made it more useful. These are the product lead's own observations; the video does not provide usage data.

Terms in this piece · Glossary
  • AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
Read the source www.youtube.com
More from OpenAI
Recommended reads
Comments

Checking sign-in…

Loading comments…