Detecting and reducing scheming in AI models - OpenAI

Scoresby

Reminded me of https://stacker.news/items/1215530/r/optimism

Remember the statement from the former employee in https://stacker.news/items/1040555/r/optimism where he said:

> We'd pay close attention to what was happening at Meta, Google, and Anthropic–and I'm sure they were all doing the same.

So I guess that thing caught their attention and staying true to the role of panicky startup that is seeing some speedbumps, they imitate. Lol