OpenAI Admits ChatGPT Work Launch Flaws, Promises Fixes

OpenAI admits it "didn't get everything quite right" with its recent ChatGPT Work launch. Users reported quickly hitting usage limits, a confusing desktop app, and problems with multi-agent workflows. OpenAI's Thibault Sottiaux acknowledged the issues, stating the team spent 24 hours analyzing feedback. They identified problems with high compute settings being too accessible and unclear usage limits. The desktop app overhaul also made familiar features harder to find, and some multi-agent workflows regressed. As an immediate fix, OpenAI reset usage limits for Codex and ChatGPT Work twice in one day. They are also adjusting default settings and fixing plugin issues. A larger update is planned for next week, promising restored sidebar features and clearer usage metrics. Additionally, OpenAI clarified that the coding tool Codex will not be discontinued, despite confusing messages to users. This matters because it impacts how users interact with and utilize these AI tools.