OpenAI opens up pretraining progress, proposes perplexity-based shared safety bars
gordic_aleksa · x · 2026-11-23
OpenAI researcher merettm announces that, for the sake of moving toward shared safety bars, OpenAI will provide more transparency into pretraining progress. In "Perplexity and Intelligence" he argues for common limits based on perplexity and other measurements, with progress beyond them gated on our ability to understand AI behavior.
Aleksa Gordić publicly welcomed the transparency move. Tying internal pretraining progress to measurable safety thresholds is a rare public stance from a frontier lab.
More from Models
- OpenAI opens up pretraining progress, but critics say key compute data is missing — SonglinYang4 · 2026-11-23
- 5 free uncensored GGUF models on Hugging Face: MiniMax, Qwen, Kimi, Nemotron — Promptmethus · 2026-11-23
- OpenAI admits pretraining barely improved since 2022 — but flat curves aren't flat capabilities — petrusenko_max · 2026-11-23
- Pretraining sets the floor, RL raises the ceiling: on how the two divide the work — JacquesThibs · 2026-11-23
- Google AI Studio's periodic-table model selector reveals Gemini 4 Neon: 1M context, $3/$15 pricing — patloeber · 2026-11-23
- Domingos: Log-Perplexity Fell to 0.75 Bits While Cost Rose Orders of Magnitude — pmddomingos · 2026-11-23