Interesting insight, though with MATS claiming 446 alumni in this post vs the 218 you found I suspect there’ll be some bias (eg the other profiles are no longer working in AI/AI safety and so MATS is less recognisable, or they’re senior enough in AI safety to have removed it from their profile eg to reduce cold messaging. I’d expect the former is more likely).
I do wonder what proportion of fellows who return for another fellowship elsewhere do so predominantly for the funding (as opposed to mentorship) and would benefit from a higher availability of grants.
Interesting insight, though with MATS claiming 446 alumni in this post vs the 218 you found I suspect there’ll be some bias (eg the other profiles are no longer working in AI/AI safety and so MATS is less recognisable, or they’re senior enough in AI safety to have removed it from their profile eg to reduce cold messaging. I’d expect the former is more likely).
I do wonder what proportion of fellows who return for another fellowship elsewhere do so predominantly for the funding (as opposed to mentorship) and would benefit from a higher availability of grants.