It’s particularly interesting that Anthropic came out yesterday and basically said, yeah, this stuff cannot be held right.

One can argue, convincingly perhaps, that Anthropic isn’t right and/or is marketing, but what they’re saying could be complete BS but the fact that there is doubt suggests that most people believe that no one can hold it right exists.

I’m quite pro AI, but given the radical asymmetry between the upside vs the downsides (the upside is at best maximum bliss for all existing humans, which has a finite limit, while the downside is the end of humanity which is essentially infinitely bad), our march forward in this area needs to be at least slightly more responsible than what we are doing now.