I suspect it is less about alignment specifically and more about deploying models in situations (and I guess organizations?) where it would be positively impactful. Anthropic's non-profit program from a couple of years ago came to mind when I was reading the post, but maybe I'm completely off.
I suspect it is less about alignment specifically and more about deploying models in situations (and I guess organizations?) where it would be positively impactful. Anthropic's non-profit program from a couple of years ago came to mind when I was reading the post, but maybe I'm completely off.