Astral Codex Ten

OpenPhil

9 min read

OpenPhil

Article

OpenPhil is a recurring organization in the Astral Codex Ten archive, appearing 5 times across 5 issues between February 09, 2022 and June 20, 2023. The archive places it in contexts such as “Why not just donate to GiveWell or OpenPhil or some other charity or foundation”; “donate to GiveWell or OpenPhil or some other charity”; “OpenPhil gives $10 million or something to some charity they learned about because of me”. It most often appears alongside AGI, Ajeya Cotra, Bio Anchors.

Metadata

  • Category: Organizations
  • Mention count: 5
  • Issue count: 5
  • First seen: February 09, 2022
  • Last seen: June 20, 2023

Appears In

Source Context

Recovered passages from the original issue text. When the raw archive preserved outbound links inside the source passage, they are listed directly under the quote.

So You Want To Run A Microgrants Program
February 09, 2022 · Original source
Right now AI alignment has lots of cash. If there’s a really good AI alignment charity, Open Philanthropy Project and Founders Fund and Elon Musk and Jaan Tallinn will all fight each other to throw money at it. So if a seemingly really good AI alignment charity asked me for money, I would wonder - why haven’t they gotten money from a big experienced foundation?
What’s your story for why you need a microgrants program? Why not just donate to GiveWell or OpenPhil or some other charity or foundation you respect?
(technically OpenPhil doesn’t accept individual donations, but if you break into their office and leave $1.5 million on a desk, what are they going to do?)
Biological Anchors: A Trick That Might Or Might Not Work
February 23, 2022 · Original source
The Open Philanthropy Project ("Open Phil") is a big effective altruist foundation interested in funding AI safety. It's got $20 billion, probably the majority of money in the field, so its decisions matter a lot and it’s very invested in getting things right. In 2020, it asked senior researcher Ajeya Cotra to produce a report on when human-level AI would arrive. It says the resulting document is "informal" - but it’s 169 pages long and likely to affect millions of dollars in funding, which some might describe as making it kind of formal. The report finds a 10% chance of “transformative AI” by 2031, a 50% chance by 2052, and an almost 80% chance by 2100.
Inline links: Open Philanthropy Project, https://drive.google.com/drive/u/1/folders/15ArhEPZSTYU8f012bs6ehPS6-xmhtBPP
There's a small cottage industry of summarizing the report already, eg OpenPhil CEO Holden Karnofsky's article and Alignment Newsletter editor Rohin Shah's comment. I've drawn from both for my much-inferior attempt.
Inline links: article, comment
Ajeya Cotra is a senior research analyst at OpenPhil. She's assisted by her fiancee Paul Christiano (compsci PhD, OpenAI veteran, runs an AI alignment nonprofit) and to a lesser degree by other leading lights. Although not everyone involved has formal ML training, if you care a lot about whether efforts are “establishment” or “contrarian”, this one is probably more establishment.
Impact Markets: The Annoying Details
July 15, 2022 · Original source
Disadvantage: It places a lot of the buy/don’t-buy decision in the hands of a single oracular funder. For example, suppose that OpenPhil agreed to do this, but a few years later they got a new CEO who was more conservative than the previous CEO, and only funded about 2/3 as many projects. Now the chance of getting a return on your investment has gone down by 2/3 for a reason unrelated to its fundamentals. Or suppose someone invested in a malaria project, but OpenPhil pivoted away from tropical diseases and now there was no hope of getting any return. Seems like the sort of uncertainty that would discourage people from investing.
Maybe a better way to look at this is abstracting out the investor and comparing “final oracular funder donates directly to charity” vs. “final oracular funder gives money to investor, causing charity to get funded”. Here it becomes clear that it’s the oracular funder who’s losing out on the tax benefits. But do oracular funders (eg OpenPhil, Future Fund) pay taxes at all, or benefit from tax-deductability? I’m not clear on this.
Realistically this is no worse than the current situation, where eg OpenPhil can fund someone to cure malaria and they can just not do it. If they deliberately embezzle the funds, it’s embezzlement and you can sue them. If they’re just really bad at their job, that seems like part of the risk you take in normal charity and normal investments.
Criticism Of Criticism Of Criticism
July 20, 2022 · Original source
We use dualist thinking, whereas many enlightened people have claimed that reality is non-dualist. This is just a selection and I don’t claim I’m being fair or representative. I’m having trouble expressing why I feel so uneasy about this. Part of it is emotional: I feel less enlightened than preached at. Do we really not think of the interdependence between individuals? Does this really make us less effective in some way? Or is this just a sort of stock criticism with such a storied tradition that everyone has agreed to nod their head to it and agree to do better later? Are we sure that becoming less individualist would be a better use of our energy than becoming more individualist? How did we achieve that certainty? It sure seems like more individualist countries are richer and better places to live. And that within those countries, the most individualist regions and social networks are the richest and best. Aren’t more intelligent people generally more individualist when you do the psych tests and surveys? I don’t know, if I thought it was a good use of our resources to move some direction on the individualism spectrum, I would be kind of interested in trying to figure out how to move even more towards the individuals-are-independent side than we are now. But realistically I don’t think this is a good use of our resources because I personally am not sure how to do that, I personally am not interested in singlehandedly inventing the science of shifting position on the individualism spectrum, I don’t think trying to invent this science is a great use of anyone else’s time, and if someone did claim they invented this science, it would take a lot of evidence for me to believe them. (again, I think I’m being terribly unfair here to the author, who was genuinely worried about these blindspots, but I’m using this as an example of my reaction to different types of criticism) What’s the EA equivalent of the s-ketamine critique? A Critical Review Of Open Philanthropy’s Bet On Criminal Justice Reform argues that one of a major EA funder’s programs was substantially less effective than their other programs, that this had been easy to notice for a while, and that despite this they continued the program instead of changing it or cutting their losses (for a while; they did eventually stop). Everyone in EA is very nice, so all of the responses were phrased as “thank you for this constructive criticism, but having said that I wanted to raise…” But subtracting out the usual niceness level, I get the impression it actually hit a nerve. There was some pretty deep discussion about whether the criticism was fair, whether it was excessively hostile to identifiable individuals, and (of course) nitpicking all the math. Here are some differences I noticed between the experience of reading the more specific criminal justice criticism vs. the more paradigmatic structures-and-individualism criticism: Before reading the specific criticism, I wouldn’t have been able to predict its conclusion. Was this program more effective than other programs? Less effective? But before reading the paradigmatic criticism, I could predict its conclusion pretty well. “We are all more interconnected than we think” is a typical piece of Profound Wisdom, and nobody ever says the opposite.
Inline links: A Critical Review Of Open Philanthropy’s Bet On Criminal Justice Reform
Not because EA is bad at taking criticism. The opposite: they like it too much. It almost feels like a sex thing. “Please, tell me again how naughty I’m being!” I went to an EA organization’s offices once - I think it was OpenPhil, but don’t quote me on that - and the whole place was strewn with the most critical books you can imagine - Robert Reich, Anand Giradharadas, that kind of thing. Can’t remember seeing Anti-Politics Machine but I’m sure it was there. Probably three copies per person. One for their office, one for their home library, and one for the spot under their mattress where other people would hide porn mags.
Davidson On Takeoff Speeds
June 20, 2023 · Original source
Last year I wrote about Open Philanthropy’s Biological Anchors, a math-heavy model of when AI might arrive. It calculated how fast the amount of compute available for AI training runs was increasing, how much compute a human-level AI might take, and estimated when we might get human level AI (originally ~2050; an update says ~2040)
Inline links: I wrote about Open Philanthropy’s Biological Anchors, an update
A lot of these big complicated analyses are salvos in a long-running conflict between a school of futurists based at Open Philanthropy and another school based at the Machine Intelligence Research Institute.
Davidson works at Open Philanthropy and is trying to flesh out their side of the story. Again, even his “gradual” takeoff won’t seem very gradual to us fleshbag humans; it will go from 20% of the economy to 100% in ~3 years, and reach superintelligence within a year after that9. Still, there will at least be enough time for journalists to publish a few articles saying that fears of AI taking over 100% of the economy are overblown and alarmist, which is a sort of gradual.
Inline links: 9