My thoughts on longtermism. On the one hand I lean towards short-termism for three reasons. 1. AI X risk is probably pretty low, ~5%. 2. Higher discount rates for the far off future are correct because current improvements have big increasing returns to scale on average (preventing a childhood case - ThreadSky

pmwwp.bsky.social • 36 days ago

My thoughts on longtermism.

On the one hand I lean towards short-termism for three reasons. 1. AI X risk is probably pretty low, ~5%. 2. Higher discount rates for the far off future are correct because current improvements have big increasing returns to scale on average (preventing a childhood case

Comments

moskov.goodventures.org•36 days ago

Personally I think #2 is a straw if you’re calling AI x risk “longtermism”. The risks are right now and you don’t to consider more than this generation to see it as high ROI to mitigate. The movement f’d up by mentioning many generations.

pmwwp.bsky.social•36 days ago

I think you are probably right here. I think I am treating AI X risk as a subset of longtermism when they more just have a large overlap. But I did sort of want to address both here.

moskov.goodventures.org•36 days ago

Well I don't think you're saying anything different than other longtermist critics, to be clear - I just tear my hair out at it as someone pretty concerned about AIXR and also not a longtermist. The existence of an *extra* idea tainted all the discussion.

pmwwp.bsky.social•36 days ago

I don't post because I think I'm a great original thinker, but because I'm addicted to social media. I do also think that if people don't post any common opinions, websites get drowned out in the kind of extremist post that we saw from a longtermist critic this morning.

moskov.goodventures.org•36 days ago

Yea, keep posting! I'm participating, not criticizing.

Do you mean this post from yesterday or a different one? I was so confused by this one.

https://bsky.app/profile/moskov.goodventures.org/post/3lggdf2s2wc27

pmwwp.bsky.social•36 days ago

I meant that one yes. I believe their thinking was longtermism->Musk->NIH cuts but the first connection there is incredibly tenuous. And maybe they meant something totally different.

pmwwp.bsky.social•36 days ago

of Malaria let's someone contribute to society their whole life, which lets another contribute down the line etc.) 3. To reduce AI X risk it's better to convince an AI engineer to become a doctor than vice versa.

On the other hand people on here are dicks to longtermists, who are my friends, and

pmwwp.bsky.social•36 days ago

give all sorts of unfair arguments against them. I do think we should also be taking some measures against AI X risk, like the executive order that Biden made. And which Trump just repealed. https://apnews.com/article/trump-ai-repeal-biden-executive-order-artificial-intelligence-18cb6e4ffd1ca87151d48c3a0e1ad7c1

timfduffy.com•36 days ago

As a longtermist I'm feeling so cared for right now ❤️

codytfenwick.bsky.social•35 days ago

~5% for AI x-risk doesn’t seem very low.

Comments

Posting Rules

Reply