What's causing mentorship bottlenecks?

By Peter Berggren @ 2023-08-09T14:25 (+10)

It seems to me like many organizations are trying to provide mentorship to new EAs, but are unable to mentor many of the applicants, even ones that may be qualified. This seems fairly unusual, even by the standards of (for instance) graduate schools, which tend to be just as selective as EA-focused mentorship programs, but offer more extensive mentorship.

Therefore, it seems to me like there must be some sort of bottleneck in the EA mentorship pipeline, which is confusing to me, given how much money is being directed into community building, and given how important this area is under the community-building umbrella.

What might be causing these bottlenecks, and how might we fix them?

Abby Hoskin @ 2023-08-09T18:09 (+10)

What is the bottleneck: There aren't a lot of people who are well established in things like AI Safety or AI Governance because these fields are relatively young. There are a lot of people interested in entering these fields because AI has recently become much more legibly a problem. So there are more newcomers than experienced experts. We want experienced experts to actually do the thing they're expert in, not just mentor newcomers. So there's a limit to how many new people can get mentored per year. 

What is the solution: I don't really know what kind of solution you're looking for. Spend more money to hire more experienced experts? It seems like we are doing this, but it's hard to hire good people, especially in relatively new fields. I guess it would help to have more funding if that's what you want to focus on?

Péter Drótos @ 2023-08-25T21:23 (+3)

I think the effect you describe makes sense, you can’t just grow a field arbitrarily fast.

But I think it might be useful to talk more about wheter we still think that trying to get into one of these very competitive programmes is currently still the best use of the talented people who are happy to use their career to do the most good.

Chris Leong @ 2023-08-09T17:58 (+2)

I asked and the AI Safety Memtors and Mentees Project is no longer active.

I’d love to see someone else start something similar:

https://forum.effectivealtruism.org/posts/CCXSdmkPvrLQZgyCk/announcing-ai-safety-mentors-and-mentees

lincolnq @ 2023-08-09T14:47 (+2)

Can you give some evidence/an example for "unable to mentor many of the qualified applicants"?

Will Aldred @ 2023-08-09T15:31 (+3)

One example is what Ben Garfinkel has said about the GovAI summer research fellowship:

We're currently getting far more strong applications than we have open slots. (I believe our acceptance rate for the Summer Fellowship is something like 5% and will probably keep getting lower. We now need to reject people who actually seem really promising.)