**[OpenAI 28-Day Sprint Day 11: Enhancing Codex Computing Power to Reduce Model Capacity Errors]** Beating AI Newsflash: OpenAI Product and Engineering Lead Tibo Sottiaux announced the progress of the 28-day product sprint on Day 11. No new features were introduced this time, but computing power was increased, and backend systems were adjusted in an effort to address the frequent "Selected model is at capacity" errors encountered with Codex. Previously, many users reported that despite having remaining subscription quotas, Codex would frequently display model capacity errors, causing task interruptions. Tibo stated that this adjustment has significantly reduced such errors. Even if they occasionally occur, tasks should now be able to resume execution automatically without remaining stuck indefinitely. [Original Link]