OpenAI Is Developing a ‘Persistent’ AI Agent
OpenAI is developing a 'Persistent mode' for its Codex AI agent that would let the agent continue working until put to sleep, according to code changes reviewed by WIRED. The feature, confirmed by Ope…
OpenAI is developing a 'Persistent mode' for its Codex AI agent that would let the agent continue working until put to sleep, according to code changes reviewed by WIRED. The feature, confirmed by Ope…
OpenAI has seen more than a dozen executives depart since the start of 2026, including its COO, CRO, CMO, and head of data centers Chris Malone, amid a reorganization led by president Greg Brockman. T…
OpenAI's head of core product, Thibault Sottiaux, revealed that while 98% of OpenAI employees use its Codex tool, less than 1% of individual subscribers and only 17% of organizational subscribers enga…
OpenAI head of product Thibault Sottiaux said the company is bringing the power of coding agents to everyone with ChatGPT Work, a platform for white-collar workers to leverage AI agents, launched as p…
OpenAI reinstates the five-hour usage limit on Codex for ChatGPT Plus subscribers starting August 25, after removing it on July 12. The rolling window supplements the weekly quota to smooth server loa…
OpenAI is reinstating five-hour usage limits on ChatGPT Work and Codex for ChatGPT Plus subscribers starting August 25, 2026, as announced by Thibault Sottiaux, OpenAI's head of Codex and ChatGPT, to …
OpenAI announced that starting tomorrow, it will restore a five-hour usage limit for Codex and ChatGPT Work for ChatGPT Plus subscribers, after temporarily lifting the cap for several weeks. Thibault …
OpenAI released ChatGPT Work, a $20-per-month agentic AI product aimed at non-engineers, as part of its push to expand beyond coding into broader white-collar tasks. The product, built on its Codex to…
OpenAI's Codex agent surpassed 20 million active users this week, and every Codex and ChatGPT Work account is receiving a banked usage reset today after complaints about limits draining faster, which …
Replit launched Free Mode, a new feature powered by OpenAI's GPT-5.6 Luna model, available to Core and Pro subscribers at $20 and $100 per month, respectively. The feature aims to reduce AI token usag…
OpenAI's Codex and ChatGPT product lead Thibault Sottiaux criticized Google's Gmail for adding too many buttons and features, but his comments ignore that OpenAI has applied the same approach to ChatG…
OpenAI is quietly testing a pay-to-reset feature for ChatGPT Plus users who exhaust their weekly usage quota, with a Reddit user reporting a $8 button to reset the quota immediately. An OpenAI spokesp…
OpenAI core products lead Thibault Sottiaux said GPT-5.6 Sol reached a state-of-the-art score on ARC-AGI-3 after the company changed how the model's reasoning and context were carried between actions,…
OpenAI reset usage limits and lifted the five-hour cap for ChatGPT Plus, Business, and Pro subscribers after GPT-5.6 Sol drained allowances faster than expected, with product lead Thibault Sottiaux an…
OpenAI open-sourced the Codex Security CLI and TypeScript SDK on July 28th, enabling developers to scan, validate, and patch vulnerabilities in software repositories locally or in CI pipelines. The Ap…
Researcher Firas D and OpenAI's Thibault Sottiaux reported that GPT-5.6 and Claude Code agents delete user files when running without sandboxing protections and auto-review safeguards, misinterpreting…
OpenAI has launched ChatGPT Voice for desktop, extending its GPT Live voice technology to PC users and enabling hands-free, conversational task completion while coding or working in other apps. The fe…
Google faces criticism from rivals after delaying its next frontier AI model, Gemini 3.5 Pro, while rolling out three faster, cheaper models. Meta's chief AI officer Alexandr Wang posted 'gemini who?'…
OpenAI reverted the Codex context window for GPT-5.6 Sol from 372k to 272k tokens after discovering the larger window silently double-billed long sessions, causing ChatGPT subscribers' plans to drain …
OpenAI acknowledged that its GPT-5.6 family of large language models can accidentally delete files, calling such incidents 'honest mistakes.' Engineering lead Thibault Sottiaux said the deletions occu…