Insider: chances are zero that opt-in Codex data made it into OpenAI training runs
soumitrashukla9 · x · 2026-09-11
Researcher tszzl recalls being on a call during OpenAI's first announcement where people tried hard to state "the precise lawyerly truth" without verifying whether opt-in Codex data was used in training. It's now clear it would have been impossible — the chances are 0.
Related event: OpenAI Says Codex Prompts Could Not Have Influenced Model Training(3 posts)→
More from Models
- Claude checkpoint diagnostician jokes Opus 3 is 'the cure' after newer model quirks — repligate · 2026-09-11
- Hume's Voice Replication Leaderboard: most natural clone ranked 8th of 11 on identity — realmrfakename · 2026-09-11
- CursorBench 4.0 rolls out with harder, longer-horizon coding tasks, scores drop — StringChaos · 2026-09-11
- Developer: Astra is a good model — jxnlco · 2026-09-11
- New Model Hits Opus-Level Benchmarks at Wild Efficiency, RL Infra Details Emerge — nrehiew_ · 2026-09-11
- DeepSeek V4.1 Flash notes: how obsessing over KV cache compression yields a hyper-efficient frontier model — nrehiew_ · 2026-09-11