| Respectfully, I think you're being a little uncharitable here. I'm on the periphery of the EA / Longtermist scene — I have friends and customers in that space and have studied and hung out some. I recently went to one of the online events — EAGx Virtual — so I got to see what people actually care about, are working on, the methodologies being used, etc. One way that might be helpful to think about "Effective Altruism" and "Longtermism" is that they're rather loose labels, more similar to "Democracy" than "Member of XYZ Political Party or NGO" — while there's some loose unifying threads across people who hold those views, there are very many different nuanced views about what good implementation can look like and how individual people and groups relate being effective and benevolent. Then, the people doing the best work in that space are doing some incredible things that have already had short-term high-impact results in the world like doing hyper-rigorous work on how to cure and treat and prevent diseases in the developing world, and then actually getting those things funded and checking the results. It's also very much _not_ uniformly "just use a spreadsheet" - a lot of people there take dialogue, conversation, and engagement qualitatively to ensure good is happening very seriously. For instance, a talk I attended at EAGx Virtual covered this - https://whatworkswellbeing.org/about-wellbeing/how-to-measur... - I haven't been able to dive deeply into the methodology, but it looks incredibly thoughtful and with many practical applications. I'm, personally, really happy people are doing work like that. Then, simultaneously there are people in the EA/Longtermist space that are working on more speculative potential problems — pandemic preparedness, reducing the risk of nuclear war, and yes AI Safety. To the latter point, empirically a lot of people in AI Safety seem more worried about very practical can-definitely-be-anticipated risks of things like, e.g., USA-China rivalry and what types of regulations, markets, and general consensus might prevent that from going badly. Obviously highly speculative futuristic ideas are sexier and get more attention than "boring routine stuff" but international conventions against, for instance, Lethal Autonomous Weapons (LAWs) seems like a really prudent thing to at least explore and see if an international consensus could be built around similar to the Geneva or Hague conventions. As for "EA-think" I don't agree with Aaronson's assessment. I'm n=1, but there's not a single monolithic viewpoint. Some really exceptional work is being done by some very thoughtful people there. Undoubtedly some bad or sloppy work is being done, as is true in just about any domain. But on balance I've found the people there to be quite thoughtful and doing some really valuable things for the world with a lot of sincerity and rigor. |