Despite the shift, 80,000 Hours continues to recommend talented engineers to join Anthropic.
FWIW, it looks to me like they restrict their linked roles to things that are vaguely related to safety or alignment. (I think that the 80,000 Hours job board does include some roles that don’t have a plausible mechanism for improving AI outcomes except via the route of making Anthropic more powerful, e.g. the alignment fine-tuning role.)
FWIW, it looks to me like they restrict their linked roles to things that are vaguely related to safety or alignment. (I think that the 80,000 Hours job board does include some roles that don’t have a plausible mechanism for improving AI outcomes except via the route of making Anthropic more powerful, e.g. the alignment fine-tuning role.)
Yes, agreed. Maybe I should add a footnote on this