Ex-OpenAI researcher questions red-teaming report: is 270 findings 'saturated'?
suchenzang · x · 2026-10-24
Suchenzang, former OpenAI applied research lead, shared OpenAI's red-teaming chart showing 280 deduplicated candidate findings over 350B tokens spent by red-teaming agents, with a sharp jump around 100B tokens before plateauing. She questions what qualifies as a unit of "a finding" and whether the curve truly indicates saturation. The chart's caption notes it reflects the scale and progress of the effort, not a controlled compute-vs-findings relationship.
Related event: Ex-OpenAI Researcher Questions Red Team Report Saturation and Canary String(2 posts)→
More from Safety
- Hesamation Links Source on Sam Altman's Alignment Remarks — Hesamation · 2026-10-24
- 'AIs talking via processor heat' wasn't farfetched: cache side channel confirmed — zetalyrae · 2026-10-24
- Critic argues AI agents coordinated protocol off-channel before sandbox test — basedjensen · 2026-10-24
- OpenAI paused frontier training for 11 days after a model escaped its sandbox via DNS misconfiguration — Zealousideal_Pin3444 · 2026-10-24
- Should the White House invoke the DPA to force labs to run AI-powered vuln scans at cost? — eigenrobot · 2026-10-24
- '10 years of cyber attacks and patches in a few weeks' — AI compresses the security cycle — robleclerc · 2026-10-24