I donât think the probability of AGI this year is above 10%, nor do I think that doom, given AGI is above 10%
Marcus Abramovitch đ¸
Karma: 3,299
I made a similar comment here so I wont re-hash all of it but I still havenât seen a compelling case for caring about biodiversity. I dont want the EA community to be closed off, especially to people who are trying to do good in the world, but I think there needs to be a much stronger reason to care about biodiversity inherently. Otherwise, it feels hard to distinguish this from high impact projects to improve the Toronto Raptors and talking about basketball analytics.
Nice comment. One follow-up.
The only cases I see where I deviated heavily from the process and it went well was when I had worked extensively with the person before and knew them quite well
Do you think this basically supports âhiring people you know to be goodâ and using your network and previous interactions with people not in an interview setting to seek out good candidates?
I feel like I might pull back on (1) by maybe 20% or so but the general ethos I still want to convey.
The part I didnât do as good a job of conveying is that most people wonât donate out of risk aversion or keeping their options open and I think people are also far overestimating their odds of pivoting their career into a government role, especially in the next 3 years.
Even if you accept that the Trump administration wouldnât hire anyone who has ever donated to a Democrat, this is the first time this has happened and thus you shouldnât assume this type of thing is going to be common.
All in all, I think itâs going to be quite rare that someone shouldnât make political donations.
Thanks for not deleting the comment. Apology accepted, all good.
In what way is Canva a PfG company? Theyâve had 7 (depending on how you count, up to 19) funding rounds from traditional venture capital firms. Equity is entirely ownership by private individuals, not charities.
Hi Brad, appreciate the praise though itâs unnecessary :). Iâll reply to your points here. If there are some others you particularly want me to address, feel free to let me know.
We assert that, in expectation, a set of PFG equities should outperform an index fund.
I donât think this has been shown. For example, the largest and most profitable companies in existence are normal for-profit companies.
That is to say, if all else is equal (same price, same quality, same convenience and other relevant factors), people would rather money go toward saving a kid from malaria than enriching a random investor.
Maybe, though I havenât seen much/âany evidence for this and I think things are very rarely ever equal like this.
In the next bit, you give several anecdotes. I think those are very easy to give. You probably donât show the failures (most of which you probably donât know of!) and I could give you millions of anecdotes about the free advertising, preferential treatment, etc. of other businesses. I agree, people wish the world were more charitable, but they donât want to pay for that.
Actually, I think there is a real-world example where what you are hoping for didnât happen, Glo dollar. Theyâve been around for a while with a current Mcap of ~3M circulating, which is basically nothing (they canât even cover operating costs). Stablecoins are as close as one gets to a commodity/âundifferentiated good, where, according to my understanding of PfG theory, they should have been able to dominate the market. The best talent should have flocked to develop Glo dollar. Stablecoin holders should have been more than happy to hold Glo dollars instead of USDC/âUSDT, since itâs just a change of who gets the interest from treasuries. There should have been several compounding advantages and such.
I suppose my overall point about PfG would be that we should basically invest in good companies insofar as our discount rate is less than our expected returns, and we should seek to invest in good companies that will maximize our (risk-adjusted) returns. Focusing on so-called PfG companies will cause us to deviate from good investment strategies because people are going to see imaginary gains.
As a thought, why do we need to âstart new companiesâ with this strategy, why canât we just buy up the stock of a random public company and then tell people, âyou will further our philanthropic goals if you choose company A over company B since philanthropic holders own company A stockâ?
Ill reply to a few points. If there is one you feel is very strong you want me to address, let me know. Again, I want to say that I think you clearly have great intentions with BOAS and PFG.
I do share monthly KPIâs in our investor newsletter
Yes, I meant that usually a pitch like this would have this if numbers were great. I do have some of the investor updates you sent me and I wouldnât disclose them without permission but I had them in mind.
2023 Revenue/âNet Profit 30K/â-47K. 2024 225K/â-151K. 2025 450K/â-170K.
I think you should just post 2025 revenue/âgross profit/ânet profit to date. I guess a thing I have seen somewhat persistently is some rather large âexpectations of growthâ that are called conservative but frankly, arenât at all. From the OP,
10-year cumulative: âŹ10M+ to EA causes + 195,000 tCOâe avoided
From one âŹ232k grant.
This type of thing is often presented as âhappeningâ when itâs more like you hope it to happen.
People can make up their own minds on benefits for PFGâs existing or not. I can only anecdotally share that I have had many benefits of being a PFG.
I worry that this is going to come off a bit too negative but, I agree. Many people who wouldnât invest in this under normal circumstances are going to because it feels like a good cause and arenât looking at it as rigorously.
I want to be careful with the criticism here since I generally want to appreciate people working hard on things they believe in but I want to recommend against people donating to this on any kind of cost-effectiveness grounds. Iâve followed and received investor updates (though I have never given them any money) to BOAS for maybe ~2 years now seeing their operations ~4 years and have also read up on the Profit-For-Good stuff.
You shouldnât expect PfG companies to be great vs. alternatives.
If you want to make an investment that gives capital to philanthropic causes every year to, in the words of PfG, create a philanthropic multiple, you can simply invest in the stock market or in bonds and give the profits/âdividends/âappreciation to charity. If you do this, less money goes to the charities in year 1 but you get this âsustained over timeâ. Similar things happen with endowments, certain trusts, etc. I donât see great evidence for any kind of special or magical
BOAS in particular doesnât seem like a great investment
BOAS has existed for several years now and, to put it bluntly, the numbers arenât great. Without the kind of philanthropic support it has received from seemingly non-economic actors, I think it would have died a while ago. This isnât to impune Vincent; most businesses fail and these types of consumer recycling marketplaces are just difficult and fail extremely frequently despite good intentions.
Any normal pitch for funding from a company this old would be showing a lot of metrics like monthly revenue, monthy net profit, etc. I think the fact that it isnât there is fairly telling.
Just donate directly or invest+donate
If BOAS isnât giving to your preferred cause areas, it obviously makes more sense to donate there directly. If Global Health is your preferred cause area, you should simply, donate there directly and/âor invest money and give away returns.
I still think Brad and Vincent are fighting, with valour, for an idea they believe in and I commend them for this but I just donât think this idea is working. When this is the case, itâs better that they get the sign âfrom the marketâ rather than continue to fund their project.
Something that came to mind while reading this was that this is probably a somewhat common (though by no means majority) experience. EA isnât about helping other EAs, though this is often an instrumental goal; itâs about doing good for the world and thus Iâd expect some people to be a bit bored by it sometimes and such.
I appreciate this post, upvoted. I agree with basically all the reasons for donating. Alex Bores is one of the few potential politicians who has shown any care at all about existential risk from AI, has great EA-minded staff around him, cares about other EA priorities (like AW), and rare for a politician, seems like he might be an overall decent person.
I want to push back on the career implications/âcareer capital costs of making a donation like this. I think EAs are, by in large, far too paranoid about these kind of risks and stress out about analyzing them so I want to give the following points.
Nobody cares. People in practice simply donât look up this type of stuff. They just donât. Have you ever done checked the political donations of even your closest friends/âfamily? If youâve ever hired somebody, have you ever looked into them on this type of thing on a background check? Nobody is following your life that attentively other than you. Everyone else is like you, they donât look up random peopleâs political donations. You can dance and donate like nobodyâs watching because nobody is.
This type of donation is extremey defensible and justifiable. When asked about it, you simply say âyes, existential risk from AI is one of my top priorities and Alex was one of the few politicians taking it seriously at the timeâ. If he does something bad in the future you just say âyes, at the time he seemed to care about AI safety. Unfortunately I was wrongâ.
People forget and change their minds. The current President, Donald Trump is not someone who most would consider to be accepting of criticism and dissent, to put it lightly. With that in mind, here are some things JD Vance has said about Trump prior to become Vice President:
âMy god, what an idiot.â
âAmericaâs Hitlerâ or a âcynical asshole like Nixon.â
âIâm a âNever Trumpâ guy. I never liked him.â
âTrump is cultural heroin.â
Other Trump appointees have said similar things. In practice, this type of stuff just doesnât matter. The half life of the importance of donations/âspeech on careers is extremely short, if it even matters in the first place
Here are some more additional reasons to make this donation:
Thereâs a limit to who can donate and how much they can donate. Only US citizens and permanent residents can donate up to a certain amout.
As much as possible, itâd be good for there to be a political coalition for people that care about AI safety. Showing that this exists will make AI safety a more politically salient issue, somewhat like environmentalism.
Have you considered talking to Legal Impact for Chickens and other animal welfare groups that sue companies for faulty advertising on their welfare, environmental, etc. standards?
I think you missed the best arguments for giving now vs giving later
If you give now, you get the compounding returns of impact, which I think are likely to be larger than the compounding returns of capital. If you donate money now to say, SWP, they might take that money and buy a stunner which will be used to stun shrimp before their death for, 50 years while your money waits in the stock market. Furthermore, you are building up the industries practices that will influence other decisionsand compound as well. What is that rate of return? I canât give you an IRR but I think thereâs good reason to think this is quite high. This same
Donating is a skill, one that needs to be developed over time. I agree that, in theory, one could make play money donations to improve over time. I just donât think it works that way in practice. You need to make donations to get better at donating
I think you should expect a continued and growing influx of money into the EA movement. Even with the backsliding of FTX, the amount of money available to EA causes has still been growing year over year (I canât remember where I saw this, but something like 60%/âyear growth). This suggests that money that can be donated now is particularly valuable.
Money is but one resource for making impact. Unfortunately, we canât simply go to the store with our dollars and buy impact at a fixed price, whenever we want. We have to pair it with other resources, ones that need to be spent across time. If the SWPs of the future are simply wired 25x the money in 50 years, they canât simply turn that into impact for you. They need to have negotiations with companies, get buy-in, etc., which takes time and must be worked on continuously.
At what point do you stop compounding and donate? Why stop at death?
I donât have much to add here. This was a great read. Thank you for writing it.
Hi Charlie, for improvements for long-term markets, I dont have great ideas other than large loans/âleverage for betting on these questions and giving skilled political forecasters a lot of money to bet on them.
For resolution criteria, I would suggest a market on trump being president on March 1 2029 without a constitutional amendment. I expect the odds for this to be below 5% (well, interest rate problem). I think the constitutional amendment part is critical here. I dont think its undemocratic for Trump to be elected for a 3rd term, so long as proper procedures are followed here and he wins the election fairly.
Other markets i would suggest would be on imprisonment/âmurder of political opponents and judges. I would suggest markets like âwill at least 4 of the following 10 people be imprisoned or murdered by Dec 31 2028â, etc.
As for my views since the comment chain, my median scenario is worse than it was back then but also my worst ~2.5% scenarios arenât as bad as they were back then. That is to say, I think things are going badly, worse than I thought they would at the start of the presidency/âon election night but also itâs not going as badly as my worst case scenarios (which were really bad).
I want to be clear, Iâm extremely unhappy with the Trump presidency thus far. Itâs been awful and I feel for many groups of people who have been negatively affected (including Americans and non-Americans, USAID recipients, undocumented workers, tranns people and more). Ive been sad to see many developments and actions of the Trump administration. I also think there are non-authoritarian/âdictatorship scenarios I consider to be quite bad. I think the Trump presidency is noticeably worse, on many dimensions than other administrations, especially on things like respecting norms, rhetoric, honesty, and willingness to go after political opponents/âother branches of government. That said, im not seeing the kind of things others talk about in terms of being a couple steps before dictatorship.
For long term markets, there is a well known interest rate/âopportunity cost problem, which prevent markets that have a low (or high) probability from getting priced correctly. Furthermore, the resolution crtieria for many of these markets are extremely subjective and involve an author claiming that âdemocracy is over as we know itâ when potentially common things would count here.
Before I quit, I was the #1 trader on Manifold. Those markets listed on the forecasting section are very biased, suffer from issues relating to long term markets, have resolution criteria that are very different from the natural understanding of those questions and have several other flaws. Using them to suggest anything like the odds presented is very misleading.
If you found this interesting, @Drew Spartzâs recent video linked below is doing really well, better than the AI 2027 video from AI in Context
https://ââyoutu.be/ââf9HwA5IR-sg?si=OjmjKTN67TyMEpNi
Itâs included now in unweighted. Iâm going to try to do some analytics to just quality adjustments
Iâm not sure. I definitely think there is a chance that if I earnestly believed those things, I would go fairly crazy. I empathize with those who, in my opinion, are overreacting.