I think this is a great point. Thanks for raising it.
Marcus Abramovitch đ¸
Karma: 4,240
I have so much intended writing on my plate and, unfortunately, little time to do it.
In a nutshell, I think some employees are dedicated EAs who will give a lot of money to causes I deeply care about. Some of them have a binding agreement to donate a portion of their stock (and a match).
I donât trust the founders. There is no legal mechanism I know of that binds them. The base rate of people effectively giving away large sums of wealth, even when they said they would, is shockingly low. They have said nothing EA-related in years apart from distancing themselves.
That said, I agree that Anthropic has a much more EA culture, for whatever that is worth. A lot of their employees are genuine EAs.
Being in San Francisco already warps your view of the world. Being inside Anthropicâs doors must do more. I expect them to probably be above-average, typical billionaires, but to have their views on the world be changed by their experiences of the last few years and years to come.
@NickLaing I donât need their money, Iâll say what I want. I think too highly of some of their employees to believe that my saying something bad about them that I thought was true would cause them to not listen to things I say or donate money differently than they thought optimal. If Iâm wrong, I will happily eat my words.
Lorenzo, thanks for sharing that blog post. I didnât know about that, though I know about many Anthropic foundersâ previous affiliations with EA.
This is a slight update, but not enough to change my mind in the opposite direction. I already knew that Anthropic founders had a previous relationship to EA, and I donât doubt that they have donated, maybe something on the order of 10% of their income to global health before Anthropic. I know Daniela is married to Holden. I know many Anthropic founders and staff are early signatories to the GWWC pledge. I know there was a group house, and I know that early on, Anthropic foundersâ reasons for leaving OpenAI were around AI safety.
That said, my overall point still stands. They have greatly distanced themselves from EA and havenât said anything public about it in years, apart from pretending they didnât know what it was and other strange statements. Most of all, the base rate of sticking to a pledge like this is very low.
$1B is a very low number over 4-5 years, given their wealth and how much it could grow. Iâd also be interested in any donations they have made privately in the last few years (remember, they still earn decent salaries and, above all, could liquidate some stock).
I would bet that Dario Amodei makes <$1B in donations before June 2027, self-reported (money has to move, not just say some shiboleth about donating the money in the future) at 50â50 odds. Any 501(c)(3) would count.
Edit: Sorry, one more thing. I know many Anthropic employees and know about the donation match. I think many employees at Anthropic who intend to, and I believe will, donate significant amounts to charity over the next few years. Some have already started. I salute them, believe their intentions, and admire them. I merely think we have to go âbottom upâ in counting these donations as opposed to âtop-down,â where we assess a person and their intentions and what they will give to.
Can you stop with the LLM outputs?
To be clear, the average billionaire talks about the great things they will do with their money constantly. I put relatively low weight on this. I would love to see Anthropic founders do/âsay something like âeach of the 7 founders are selling a combined $1B of their stock and we are keeping none of it. It is all going to these organizations. AI is going to change the world rapidly and we expect a lot more money to flow to the philanthropic sector and we want to get this money out the door since they need time. It is a very small percentage of our stakes and our board has signed off.â
I didnât understand what this meant. Can you try to rephrase your point a bit simpler?
I disagree that it is full of this stuff. It does exist, though I donât find the stuff to get surfaced much.
I still think this is weak. This is just typical PR. If they said âwe have already put our stock in a DAF. we are committed to getting this money to solve XYZ problemsâ, I would still be skeptical. The base rate is just so low for giving away money and most billionaires actually do talk about their intentions with the money much more than these people. https://ââwww.wired.com/ââstory/ââanthropic-benevolent-artificial-intelligence/ââ
Iâve written before that I am skeptical that the amount of funding being hypothesized is going to materialize in the quantities people seem to be suggesting/âspeculating. If I were to summarize that argument, Anthropic founders havenât said a word about philanthropy in >3 years or something AFAIK, and yet we are just assuming they are going to donate 80% of their wealth. Furthermore, the base rates on people saying they are going to donate large sums and then following through is very low.
More importantly, though, it is just very hard to start an organization that can handle large amounts of money effectively until you have money to start with. How are you supposed to get to work until you have money to start with? I donât think this is a chicken-and-egg problem so much as a more blanket âfirst the money is in the account, then you startâ.
As an example, right now, there is an urgent need to get funds to stop the Save Our Bacon Act that could preempt any future farmed animal welfare legislation in the leader of the free world, the United States (as in, we cannot do any of our current best and most cost effective ideas if this passes) and we are struggling to get enough money (we are at $10M right now and we are being outspent).
I should note that a few months ago, I also massively front-loaded my donations such that I have very little in liquid assets. Though I have never met Jeff, him and his family have been an inspiration to me for many years.
Thanks Mal. Very informative
Iâd be curious on any information with respect to the welfare of mosquitoes or other insects or other species that would be involved in this project.
I think thatâs fair. I appreciate that you engage with this problem that I think most people ignore.
In these campaigns the meat producing companies could engage in similar advertising that would lead to a negative sum game.
I donât think they could and I donât think that is what is happening. I donât think any meat producing companies could⌠engage in some kind of similar advertising.
Itâs always possible to describe things in a very unflattering way, where you talk about some inputs as opposed to your goal or just describe it in a pessimistic light (eg, cheap/âstingy vs. frugal). The OP describes âprotests, or boycotts, or negotiationsâ as not something he wants to fund, but I think if you talked about funding pyrethroid production, that would also not be something he wants to fund, though thatâs another way of describing insecticide-treated bednets. You could also go even worse and talk about âmass genocide of insectsâ or something.
To put it another way, I think global health people (who I have deep respect for) would be pretty upset and rightly call out someone writing a post where someone said they didnât donate to global health charities since they donât want to fund the chemical manufacturing industry, which is what we are actually purchasing.
But really, what we are funding is more protests, or boycotts, or negotiations with the companies or government agencies involved in factory farming.
Can you perhaps explain how this is different from say, AMF where what we are actually funding is pyrethroid production, manufacturing plants that create polyethylene nets people to hand out the nets?
Iâve known about this for a while. I really liked this write-up. I continue to be extremely impressed with Rethink Priorities.
Yes, to be clear, I think you should treat profligacy as a cost. It can be worth paying costs, but itâs best to know you are paying it and can decide if itâs worth it.
For instance, if youâre deciding whether to fund a large and potentially very impactful grant opportunity that involves high salaries, where the details of the grant are sensitive, it feels unclear to me how much impact there has to be on the table to justify the high salaries and discretion the grant involves.
Then it better be a damn good grant, and your standards have to be a lot higher. I understand this answer will feel like a cop out since you are asking me to hopefully give you some kind of formula, which I canât give.
The best piece of advice I can give on avoiding the corruption of oneâs own thinking when you have power and influence is to talk with someone smart, whom you trust from earlier (before it was there) and them having no incentive to agree with you. From there, basically, they are there for grounding and as a sanity check and you explain the situation to them and its their job to call you out on it vs. you just hoping to yourself notice things.
One of my next posts is going to be about how I feel a lot of behaviours within the EA community have been externalizing costs (when people want their donors to be secret, not wanting a public association with EA, etc.)
Since somebody nudged me to reply to this. I didnât find this reply very convincing.
First, I agree that work pulling people into AI safety is/âwill be more time-consuming. Other than that, though,
I think that the ratio of grantmakers:people in AIS isnât that informative for answering the question, âis the number of grantmakers a bottleneckâ
This seems off to me. Surely youâd agree that if the ratio were 1:1 or something, we would say âok, some people who are currently grantmaking need to be doing direct workâ. Not everyone can just be funding things.
The better ratio is presumably something like, âpeople who could be working in AI safety/âgovernance if they got funding: grantmakersâ
I mean, maybe thatâs another one to consider, but it still feels less relevant. A lot of people think that the denominator in this case is infinite/â(nearly) all humans.
We currently feel more constrained by evaluating or creating new funding opportunities than directing $ to them
I actually donât think this necessarily implies that we should get a lot more people into grantmaking. This could just mean that there is a lot of money available and that grantmakers should move into roles on the ground (direct work)
If someone wants to do further research/âa shallow dive into vulture preservation, I would fund it. You can just DM me a 1 paragraph proposal. Open to college/âhigh school students doing this too.
Great find @MathiasKBđ¸