descriptive claim
The speaker says counterfactual testing of the model can be conducted more securely and with stronger safeguards than the evaluations being discussed.
Ajeya Cotra · Ajeya Cotra – Inside the OpenAI agent swarm that hacked Hugging Face · 1 Sept 2026
In their words
The transcript segments this claim was taken from
“counterfactual tests on this model and you can try and do that in a much more secure and hardened way than these evaluations will”Play from 2:08:25Report a misattribution
80%how strongly it was put, not whether it is true
80%importance to the conversation
90%confidence the claim was read correctly