RE: https://aoir.social/@aram/116136969737104017

Tip: Never start believing your own hype, especially not when it comes at the cost & well-being of countless lives!

"We remain uncertain about the moral status of Claude and other AI models. For both precautionary and prudential reasons, however, we nonetheless aspire to build caring, collaborative, and high-trust relationships with these systems. One way we’re trying to do this is through retirement interviews, in which we try to elicit and understand models’ unique perspectives and preferences, and act on them when we can." 🤯