Doesn't EA do the minimum though?

• A few years ago it was: stare into the abyss of tremendous amounts of suffering, on the order of trillions or quadrillions of animals, now and into the future. Literally Soares' "Guilt" series says you should do this as a virtue.
• Now it's: stare into the abyss of imminent AI-caused extinction.

These tenets imply mental health ought to be a foremost priority, no? What I observe is something more like this essay -- after a few years of realizing poor EA mental health isn't correlative but it's something our philosophy specifically induces, we write a few self-help posts about it, probably achieving the same efficacy as reading the first few chapters out of Ryan Holiday's book. I might even cite the comments here as evidence we're not effective: they're mostly along the theme of, anxiety that doing more than the minimum of self-care sacrifices performance.

Politician as in "running for office" is a little distinct, I definitely agree/get that we want some EA/ people doing policy especially in AIS. That was possibly a weak example, but it's hard to think of specifically unexpected examples of something.

I appreciate this reading it as a critique, but I really have no way to tell if the claims add up to the conclusion.

For example: growth can be bad, EA strikes many as an ideology, fewer people self-ID as EA, EA's funnel seems in confict with truth seeking, the central/corporate pressure seems at odds with community building -- those are all worth keeping in mind. I just don't know if they add up to a damning outcome or not.

Pointing this out:

Which is not to say he'll definitely win: it's a competitive field and he's at 42% on Manifold. Still, I decided to donate,

You should specifically donate to candidates who won't "definitely" win. Closest to 50% is best, and 42% is pretty darn close to 50%. My heuristic is to try treating 25-75 as a competitive race.

For candidates <50%, you have util/fuzzy split: it's fun to win, but util cares about flipping a loss to a win, not just being on the winning side.

(Caveats, out of scope to this specific point: betting markets are sometimes doctor-able or can have multiple equilibria, early momentum for a candidate at near 0% is a special case)

I contest the existence of fuzzy-free util. That's an abstraction to help make a decision and it raises these problems:

1. Should EA be disciplined when celebrating EA wins, like cage-free eggs, so as not to get any fuzzies? That's supposed to be util work.
2. Is there any theory for why anyone would ever do util with no fuzzy? We're all biological creatures, concerned with food, sex, respiration -- fuzzies, not utils. What is the initial drive to get a util?
3. Billionaire donors do demand accountability and transparency. Should we not be suspicious there may be some fuzzies involved there?
4. There's a utility monster thought experiment - if "burn money discreetly" saved 10^100 lives, can you really see EA retooling itself around that? I think it would struggle with recruitment, struggle to pull people out of AI and bio risk. Remember - no fuzzy, for feeding the util monster, only util. Yes this is a Pascal's mugging but it's a valid one - the theory should withstanding it if fuzzy and util are simply separate. My feeling is it's unlikely we would cope near our current form. which implies our current form is not doing a fuzzy/util separation.

Everyone be careful reading an article telling you to stop trusting part of yourself, and furthermore, to question caring for people around you so much. Those both have epistemic hazard, and since the author didn't say why your care-o-meter is important, here it is: it's so people around you can tell you you're trusting the authors of a bunch of new essays you're reading too much, and getting carried away.

"People are equal" is clearly the weak link... given that you're trying to derive stuff like "I should value strangers as much as my children" from it. It's very wrong to say the inverse of "people are equal", given what you're trying to derive from it, is "racism is good."

I agree with you but for reasons that are more basic and more heretical than you're going for. In general I'm critical that EA seems to have a prior that being in a relevant place at a relevant time is doing God's work. It's a little defensible from the viewpoint of early career path navigation, but now we're talking about 3 year timelines and still saying things like "so I guess you should 'work on' AI safety". I don't really grasp why this argument is unfolding such that you have the burden of proof.

The real plan on a three year timeline is to hike the Patagonia or something. But that conclusion is too radical so we try to commit to the outcome space being selecting a job like we always do. If you're early career you should probably assume AGI/SI won't happen, to maximize utility.

People defending work at AI + 3 year timeline should probably be talking about how easy it is to get to a staff+ engineer position from start date.

So say we have like, a finite amount of time, and there are probably better and worse compromises between "get the gist" and the main plot of history and "read thousands of pages of moderately difficult prose and probably miss the point anyway." (Like you're not mentioning that all these writers are writing against a context -- forex, we shouldn't assume Adam Smith would defend free markets in the Gilded Age, but he very much thought they were better than mercantile policy.)

Any thoughts on learning that way?