This is a special post for quick takes by Jenna_Parr. Only they can create top-level comments. Comments here also appear on the Quick Takes page and All Posts page.
Has anyone looked into Accenture, Anthropic’s planned embedded evaluator? Or do they have opinions?
I had Opus look, and it said they might be notably less independent and more accommodating than orgs like METR. They sold Claude to customers before this, and are being paid directly by Anthropic to be auditors. But that’s a lazy investigation on my part, I don’t trust it much. I’m wondering if I should look harder.