I talked to a friend that works at anthropic and he said that actually it would be very hard to get out of their charity commitments. They'd have to create a good-for-nothing charity and donate to that

From @mqp on discord

some premises i have influencing this feeling (as usual, assuming that there is not some kind of near-term singularity that makes everything impossible to predict. also assuming that there is not some catastrophic AI bubble collapse wiping out people's brokerage accounts, which many people reasonably expect there to be):

1. the rich people at anthropic and openai in question may think that they feel very charitable and very interested in giving to EA-like causes, but i have no reason to believe that this feeling will survive any "change in wind", i expect them to behave increasingly over time like average rich people, and give away average amounts of money to average quality destinations at an average rich person rate

2. if there were "100 givewells" the 5th-100th givewells would end up ultra spammed by random professionals who want jobs anywhere, not especially ones who cared about effectiveness or altruism, resulting in totally average charities

3. if there were "333 arc institutes" the 5th-333rd arc institutes would end up ultra spammed by every scientist who is looking for grants from anywhere, not especially ones who had great research, resulting in totally average grants

Happening or have happened? I imagine foundations may be created and pledges made, but that's quite different than that money actually granted in any significant quantities to non-profits. Also, are you counting political donations?

@DylanRichardson
happening or happened! if it looks like people are really following up on their commitments and the commitments would be hard to get out of, that's evidence for a yes resolution. Could resolve to a high percentage if that's all that's happened.
If the political donations are largely due to charitable causes maybe I would count them? If e.g. a candidate is running on animal welfare and AI safety and that looks like the best way to achieve those things, I could be convinced

@DylanRichardson They'd have to be new commitments with fanfare tho, bc they've already committed to donate money iiuc.

wouldn't be surprised if a lot of people at anthropic don't want to do much charitable giving just yet, if they're truly AGI-pilled and think their money will have a lot more leverage sometime soon

@MingCat yeah maybe a two year horizon would be better