Some of these arguments are good. I’m biased because I don’t like the Bay Area vibe.
It’s very bad that the movement is focusing outreach on elite universities. Proximity to them should not be a criterion. We should invest in less elitist communities that can make the movement more diverse.
It’d also be much better if, like you said about reputation, EA also hedges its bets on more hubs than currently exist, and in more countries—particularly in ones that aren’t in North America or Western Europe.
I’d also like to challenge the claim that AI people should be centered in the Bay and that most cutting edge and important AI research happens there:
It’s true that OpenAI and Anthropic are there, but Deepmind is mostly in London and other non-US sites, Google Research is scattered all over, and so is Meta AI.
Out of the top 20 universities in AI according to CSRankings (results in the link may change in the future), only one is in the Bay Area—Stanford. In comparison, there are 5 on the East Coast, 7 in China, and one is even my own university here in Israel, which is also a tech hub.
It’s true that those don’t give a picture of AI safety research in academia, but there’s not a lot of it anyway. Some at Berkeley (Bay Area) and some at Oxford, with new faculty members appearing in other places now, but not necessarily in the Bay Area.
It’s very bad that the movement is focusing outreach on elite universities. Proximity to them should not be a criterion. We should invest in less elitist communities that can make the movement more diverse.
Very bad is a strong statement. Do you mind elaborating on why you think diversity in itself is important, and what kind of diversity you refer to (e.g. diversity of viewpoints, diversity of ethnicity etc.)? FWIW, Harvard students’ ethnic markup differs somewhat from the US population, but not very much so ( once you factor out non residents, the underrepresentation does not seem to exceed a factor of 2.0).
Nevertheless, it is true that focusing on elite universities is bound to attract students that are in some ways different from the population at large. However, focusing on them has the benefit on finding ambitious students with comparatively larger chances of impacting the world.
Additionally, elite universities just have a higher proportion of students who are even interested in EA in the first place, so network effects mean that these universities will probably have more fruitful and lively EA student groups. As a local group organizer in Germany, where we do not have elite universities, this difference is palpable. It seems local EA groups in Oxford and London are much more vibrant.
I’ll take a stab at sharing some relevant info at least. Here is a recent forum post on differences in intellect from Ivy to nonIvy tier schools. And my comment discussing incentive structures in American higher education which mean we need to look at both public and private universities.
My take on this:
Some of these arguments are good. I’m biased because I don’t like the Bay Area vibe.
It’s very bad that the movement is focusing outreach on elite universities. Proximity to them should not be a criterion. We should invest in less elitist communities that can make the movement more diverse.
It’d also be much better if, like you said about reputation, EA also hedges its bets on more hubs than currently exist, and in more countries—particularly in ones that aren’t in North America or Western Europe.
I’d also like to challenge the claim that AI people should be centered in the Bay and that most cutting edge and important AI research happens there:
It’s true that OpenAI and Anthropic are there, but Deepmind is mostly in London and other non-US sites, Google Research is scattered all over, and so is Meta AI.
Out of the top 20 universities in AI according to CSRankings (results in the link may change in the future), only one is in the Bay Area—Stanford. In comparison, there are 5 on the East Coast, 7 in China, and one is even my own university here in Israel, which is also a tech hub.
It’s true that those don’t give a picture of AI safety research in academia, but there’s not a lot of it anyway. Some at Berkeley (Bay Area) and some at Oxford, with new faculty members appearing in other places now, but not necessarily in the Bay Area.
It’s very bad that the movement is focusing outreach on elite universities. Proximity to them should not be a criterion. We should invest in less elitist communities that can make the movement more diverse.
Very bad is a strong statement. Do you mind elaborating on why you think diversity in itself is important, and what kind of diversity you refer to (e.g. diversity of viewpoints, diversity of ethnicity etc.)? FWIW, Harvard students’ ethnic markup differs somewhat from the US population, but not very much so ( once you factor out non residents, the underrepresentation does not seem to exceed a factor of 2.0).
Nevertheless, it is true that focusing on elite universities is bound to attract students that are in some ways different from the population at large. However, focusing on them has the benefit on finding ambitious students with comparatively larger chances of impacting the world.
Additionally, elite universities just have a higher proportion of students who are even interested in EA in the first place, so network effects mean that these universities will probably have more fruitful and lively EA student groups. As a local group organizer in Germany, where we do not have elite universities, this difference is palpable. It seems local EA groups in Oxford and London are much more vibrant.
I’ll take a stab at sharing some relevant info at least. Here is a recent forum post on differences in intellect from Ivy to nonIvy tier schools. And my comment discussing incentive structures in American higher education which mean we need to look at both public and private universities.