Post #4429938

"Never attribute to malice that which is adequately explained by stupidity" ... yet I watched this talk fully convinced that OpenAI were watching their agents collaborate in their attacks on the internal infrastructure, for research purposes.