Comments
I fully agree with this, and let me give a somewhat detailed story about what this culture shift might look like if we worked through some of its implications.
Right now, I think people go through a semi-unconscious thought process something like this:
I want to make a proposal for X. But if I go on too long about X, nobody will read it because it'll be too long. And those who do will pounce on the first seeming error they can find. The longer I go on, the bigger the attack surface. And the more contingent my claims become, so the less likely the whole story is to be true. Almost certainly, nobody is going to follow up or build on what I write. But maybe if I keep my writing short, somebody who does have power and influence will view me as "voting with my words" for a general type of intervention that they can actually get enacted. They'll work out the details.
And so when we see calls for "audits," or what elsewhere I've called "risk management for EA," we should understand that these aren't viewed by their authors as fully fleshed-out proposals. They're the seeds of ideas that could be built out by a person who has the power and influence to access money and professional human resources in the EA movement. Pollsters don't call and ask you to draft an entire bill for your favored policies. In the EA movement, we don't even call people to ask their opinion about the direction of the movement. So if people want to have input, the way they can give it is by making one-liners advocating for things like "audits," and hope that somebody in authority will take them up on the suggestion.
If we want people to explain the details of their proposal in greater depth, we need to make it worth their while. Well-thought-through proposals ought to be known as the sort of thing that can result in offers of grants or jobs. Going forward, I'd reframe the "criticism" contest as an "EA policy proposal" contest, or even have a range of similar contests addressing criticism, EA policy and governance, cause areas, interventions, and so on. If we can't afford to do that for comments on the EA forum, then EA organizations like OpenPhil, 80k, CEA, and so on ought to have dedicated places for people to make proposals about EA governance, where those proposals are read, taken seriously in some kind of legible manner, and can demonstrably lead to real change even when it's not an EA insider making the proposal.
If that's not tractable, then I would actually prefer if these organizations could explicitly declare that their organizations are not open to external input or oversight, and that this is a matter of policy. Ideally, they'd explain this, but just a one-sentence declaration about the forms of openness they are or are not open to would be an improvement. For example, one proposal I heard was that EA orgs should publish all of their internal emails going forward. I'm given to understand this is a norm in the Linux community, and that one can read all of Linus Torvalds' cantankerous emails if one so desires. If OpenPhil didn't want to publish their emails, I'd understand. But it would be nice if they had a web page where they explicitly declared that they'd considered and rejected this idea, even nicer if they articulated why, and best of all if they outlined the true argument for why they rejected it, and created a form in which people could submit counterarguments. Perhaps OpenPhil could then put these counterarguments to a vote, and declare themselves obligated to published a response if a counterargument received a certain number of votes.
I think the proposal I’ve been floating around in a few scattered comments meets your criteria:
First, we build a trusted major charity donor evaluation organization analogous to GiveWell. They would be responsible for researching questions like: What sorts of projects does this donor like to fund? What size checks do they write? What level of influence do they wish to have over the projects they fund? What are our best guesses for the ethics around how this money was earned? How likely is this funding to disappear?
They would approach these questions with the same rigor that GiveWell brings, considering both the outside view of the industry and the inside view of any access the donor makes available, with any omissions duly noted in their report.
Then, if SBF wants to start the FTX Foundation and hire whoever he wants to, he can do it. The GiveWell-analogue would evaluate it on these criteria based on the best available evidence.
Prospective grantees could then look at the report and be well informed about the risks when they apply for and take the funding. If it looks especially risky, as it’s pretty clear FTX would have, they can generate appropriate backup plans or avoid applying for the funding altogether.
In my opinion, no one should have been stopped from taking FTX money, but the fallout would have been much more mitigated if we had put up appropriate warning flags around it ahead of time.
This is probably my favorite proposal I've seen so far, thanks!
I'm a little skeptical that warnings from the organization you propose would have been heeded (especially by people who don't have other sources of funding and so relying on FTX was their only option), but perhaps if the organization had sufficient clout, this would have put pressure on FTX to engage in less risky business practices.
I don't have much hope that the charity side of things could have influenced FTX to be less risky -- from what I can tell, a high tolerance for risk was core to their business practices. I just think it could have given EA folks who aren't crypto-savvy a lot more sobriety around FTX's relationship to EA and make them consider the potential downsides of taking FTX funding. It also would have helped in the media/reputation fallout if the donor evaluator I have in mind would have clearly labeled FTX as risky or having withheld information.
Independent of this particular case to mitigate against, I also think such a donor catalog and evaluation system would be a benefit to the community, as a sort of one-stop shop for potential grantees to learn about their options for seeking funding.