Researcher pushes back on OpenAI chat-snooping claim: 'cartoonishly wrong' and reputationally suicidal
BlackHC · x · 2026-09-08
Reacting to allegations that OpenAI accessed user chats to scoop research results (the Tristan/Sebastien dispute), BlackHC lays out a skeptical case:
- Highly unlikely: the reputational risk would be enormous, many insiders would see it, and such an action would be "cartoonishly wrong and visible"
- Key question: what account type and settings were used for Codex — did the user explicitly allow OpenAI to train on their conversations?
- One side of the story so far: waiting on Tristan's communication records; Sebastien likely didn't issue threats the way depicted
- Direction was already public: DeepMind and co-authors including Tristan posted a paper pointing that way last year
More from Companies & People
- Fields Medalist Tsimerman founds Mathematical AI Safety Institute, hiring 10-100 mathematicians — AndrewCritchPhD · 2026-09-08
- Ex-DeepMind's Danijar Hafner builds stealth startup teaching humanoids to plan in unseen rooms — nordicinst · 2026-09-08
- Palantir names Nebius preferred partner for sovereign AI compute stack — BrettKrieger12 · 2026-09-08
- Zvi Mowshowitz on Magic, Poker, Jane Street Trading and His AI p(doom) — TheZvi · 2026-09-08
- NVIDIA details how Domyn specializes Nemotron models for regulated industries — NVIDIAAI · 2026-09-08
- After GPT-6 Astra, Brockman Hedged on AGI While Jensen Huang Declared the Era Has Begun — Alec_Coughlin · 2026-09-08