Yeah that’s basically it. Money they get is ONLY spent on defeating this bill.
Bentham's Bulldog
Yes it was stripped from the earliest senate version but might be added back.
Post Anthropic funding we’ll probably have about 50 billion dollars a year. Now, that’s not enough to sponsor every conceivable project that will make the world better, but it is enough to sponsor the main portfolio of the highest-impact projects.
Even if future people matter .001% as much as present people, the argument for SL goes through.
Also, it just seems bizarre that the moral importance of a person would depend on when in time they exist. If someone time-traveled back from the future to the present, would it start being 100x worse for them to stub their toe?
I’m not sure I quite get what you’re saying. The point is just that if actions have positive EV and impact on the order of 10^50 lives potentially, that is very likely to swamp other positive actions.
So cool!
Awesome post!
Note the U.S. hasn’t had 10+% GDP growth since the great depression. But yeah I’d be happy to take some bets about this—north of 5% for instance.
Weirdly aggressive reply.
First of all, the AI 2027 people disagree about the numbers. Lifland’s median is nearer to 2031. I have a good amount of uncertainty, so I wouldn’t be shocked if, say, we don’t get the intelligence explosion for a decadeish.
“you’ve predicted a 95-trillion-fold increase in AI research capacity under a ‘conservative scenario.’” is false. I was just giving that as an example of the rapid exponential growth.
So the answer, in short, is that I’m not very confident in extremely rapid growth within the next few years. I’d probably put +10% GDP growth by 2029 below 50%.
Sure. I’d bet that in the next 15 years the U.S. will have 10+% GDP growth in at least one year.
Less sure about pre 2028 bets.
I meant why the low probability of bee sentience.
Out of curiosity, why so low?
Want to come on the podcast and argue about the person-affecting view?
Probably our disagreements are too vast to settle much in a comment.
I mean, that might help with a few problems, but doesn’t help with a lot of the problems. Also, it just seems so crazy. Giving up axiology to hold on to a not even very widely shared intuition? Giving up the idea that the world would be better if it had lots of extra happy people and every existing person was a million times better?
Amazing stuff!
The apples being unbounded thing was just a brief intuition pump. It wasn’t really connected to the other stuff.
I don’t think the argument actually requires that different value systems can be compared in fungible units. You can just compare stuff that is, in one value system, clearly better than something in another value system. So, assume you have a credence of .5 in fanaticism and of .5 in bounded views. Well, creating 10,000 happy people given bounded views is less good than creating 10 trillion suffering people given unbounded views. But that’s less good than a one in googol chance of creating infinite people given unbounded views. So by transitivity, a 1/googol chance of infinite people given unbounded views wins out.
Awesome post!
Would be curious why people are downvoting.
Yes.