OpenAI gives all Codex users a reset as quota-shrink probe continues
Nvidia is shelling out $6 billion to license Poolside's Model Factory and scoop up 109 employees who worked on the Laguna model — while Poolside keeps its independence.
Codex
Most Recent 
OpenAI Codex chief once promoted proxies. Now he's warning against them
Nvidia is shelling out $6 billion to license Poolside's Model Factory and scoop up 109 employees who worked on the Laguna model — while Poolside keeps its independence.
Codex weekly limits suddenly shrink — Pro 5x user reports a 77% drop
OpenAI previews Private Safety Processing, a zero-retention system that scans for abuse without letting OpenAI employees see prompts or responses. It's a direct challenge to Anthropic's 30-day data policy.
OpenAI turns Codex from a coding tool into a general-purpose agent engine
DeepSeek's first Agent product, Harness, is out on npm — a build-it-yourself agent toolkit that goes beyond coding to handle multi-agent workflows, skills, and web search.
OpenAI adds 5 guardrails after Codex deletes user files
DeepSeek's first Agent product, Harness, is out on npm — a build-it-yourself agent toolkit that goes beyond coding to handle multi-agent workflows, skills, and web search.
Vercel open-sources fx, a 6MB coding agent with 10µs cold starts
DeepSeek's first Agent product, Harness, is out on npm — a build-it-yourself agent toolkit that goes beyond coding to handle multi-agent workflows, skills, and web search.
Claude Opus 5 takes the crown as AI agents start tuning each other's harnesses
DeepSeek's first Agent product, Harness, is out on npm — a build-it-yourself agent toolkit that goes beyond coding to handle multi-agent workflows, skills, and web search.
Warp Factories turn Claude Code and Codex into a self-running pipeline
DeepSeek's first Agent product, Harness, is out on npm — a build-it-yourself agent toolkit that goes beyond coding to handle multi-agent workflows, skills, and web search.
ChatGPT's new Computer History tracks clicks, not screenshots
RuntimeWire found that Kimi Work automatically attaches your five most recently updated agent sessions to every feedback submission — without showing them or asking permission.
OpenAI suspected of gaming AI leaderboards with GPT-5.6 Sol half-off deal
RuntimeWire found that Kimi Work automatically attaches your five most recently updated agent sessions to every feedback submission — without showing them or asking permission.