I've been interested in Effective Altruism for a long time now, but I've never been comfortable with a lot of the EA dogma.
In particular, I'm very skeptical about how much we can reason about the far future; I think there is just too much uncertainty about what the world will look like in 100 years to try to make some rational computation about how to effectively spend $100 today.
I'm also skeptical about how much to apply our moral intuitions to animals other than humans, and extremely skeptical of attempts to compare animal suffering and human suffering to come up with some sort of calculus of "X insect lives are worth Y sheep lives are equivalent to one human life (so we should spend N dollars to improve the lives of insects instead of spending them improving the lives of sheep or humans)."
In my head I split the effective altruism movement into four quadrants, with short-term, human-focused charities being the ones I focus on.
 |
| (AI-generated image) |
GiveDirectly is still my favorite charity, because they actually measure their effectiveness and can talk to the people they help to make sure what they are doing is actually helping.
I don't think anybody really knows how to measure the effectiveness of something like AI safety efforts, so if I was King of the Effective Altruists I'd tell the AI safety people to go march under a different banner. Same for the shrimp welfare folks. I don't think it helps the EA movement to be associated with causes that can't easily show their effectiveness.
I'm not saying that those causes are bad or that somebody shouldn't be working on them-- just that "Effective" is the key part of "Effective Altruism" and it is very easy to fool ourselves about our effectiveness when we're excited and working hard on something. I remember lots of times when I worked for hours "optimizing" a piece of code just to see it run 11% slower when I benchmarked it.
Grok tells me there's active discussion/debate in some EA forums about measuring effectiveness for AI safety work. Good! I'll just say that as a lazy effective altruist who doesn't spend any time keeping track of the latest discussions... I don't see much "epistemic humility" coming from the more speculative areas that are lumped under the EA umbrella.
Posts
