OpenAI reportedly diverts 10% of compute to safety amid growing compute shortage
kimmonismus · x · 2026-10-14
- Per The Information, OpenAI is racing to make its next flagship model safe to release after canceling a less capable model over safety concerns. The successor reportedly uses loop transformers that process problems repeatedly internally, making reasoning harder to monitor, and will undergo US government testing that could add up to 30 days before release.
- The growing compute allocation to alignment may explain OpenAI's visible compute shortage: the $200 plan briefly paused new subscriptions, usage limits have tightened significantly, and GPT-6.1 remains very slow even in Fast Mode — a sign data center expansion isn't keeping pace with demand. Employees still expect one or two new models by year-end.
More from Models
- Opus 5.5 still leads a month later, but open-weight decision models are closing in — airesearch12 · 2026-10-16
- Google restricts Gemini Pro access for Google Plus subscribers, matching free tier — EmergencyLess8861 · 2026-10-16
- Meta releases Muse Spark Open, trained with vLLM-powered RL inference from day one — colinmcnamara · 2026-10-16
- Meta open-sources Muse Spark: runs on two Mac Studios, license bans passing output off as human-written — mawcopolow · 2026-10-16
- Golden Gate Claude explainer resurfaces: interpretability premise still holds — HarperSCarroll · 2026-10-16
- Hugging Face livestreams release of a new open-weight model — huggingface · 2026-10-16