I’m definitely in favour of improving the mentoring pipeline and maximising the prestige difference, although I suspect that we can’t pull away capabilities researchers at anything near the scale you’re envisioning. I do think it is possible to slow capabilities somewhat by pulling away top researchers, and some effort should go into this strategy, but I wouldn’t overinvest in this for the purposes of slowing progress since I expect we could only slow things on the order of years. That said, drawing top capabilities researchers will provide us with vital talent for alignment and also make it much easier to recruit other people.
I’m definitely in favour of improving the mentoring pipeline and maximising the prestige difference, although I suspect that we can’t pull away capabilities researchers at anything near the scale you’re envisioning. I do think it is possible to slow capabilities somewhat by pulling away top researchers, and some effort should go into this strategy, but I wouldn’t overinvest in this for the purposes of slowing progress since I expect we could only slow things on the order of years. That said, drawing top capabilities researchers will provide us with vital talent for alignment and also make it much easier to recruit other people.