We Are Always Talent/Idea Constrained
Trillions have been earned; it's you job to give them something worth giving to
I recently had a chat with some of the friendly folks at Open Philanthropy about their perpetual conundrum: how to turn money into value. Now you might respond to this by saying, “Oh, it must be easy! Just give all the money to <insert idea the speaker likes>.” To this I’d respond, “Well, have you considered that your idea, like almost all ideas, is bad? And that you don’t have amazing people to execute on it?”1
This is Frontier Analysis. Most things don’t work. Even most things people try don’t work. Even most things that do work don’t work that well. Open Phil and others, I speculate, are keeping some of their powder dry in hopes they come across things that work really well to make the world a better place.
Just look at the AI safety landscape. Open Phil funds a bunch of prosaic technical safety and alignment research or wonkish palatable-to-industry policy ideas, both grounded in today’s models, which might not even be very good model organisms for the future models we’re actually worried about. The rationalists point out that this seems like incredibly weak medicine to avoid human extinction at the hands of superintelligence, and that Open Phil is wasting its money.
For themselves, the rationalists think it’s wiser to try to wake up broad swathes of the public to the fact that AI is going to kill us all soon and that we ought to simply coordinate and shut down frontier development — an admitted long shot of a proposition whose only virtue is that it at least works conditional on getting the coordination in the first place.
The only reason either of these strategies makes any sense to anyone is because the prize at the end — avoiding human extinction and/or securing a vast, flourishing, long-term future — is big enough. The bar for the promisingness of interventions with smaller prizes like improving air quality or advancing meat alternatives is even higher. And perhaps it should be, given the relatively straightforward nature of those problems. Things like air quality and meat alternative speak especially strongly to the overall issue though, because it’s exactly in these comprehensible, low stakes places that we’ve seen so many things not work at all.
Against this backdrop, I see EAs getting excited about the prospect of securing new billionaire funders for the causes we care about. But the point above is that money doesn’t really do causes any good. Money does interventions good, by scaling them up, assuming the interventions help (a lot) in the first place. If you don’t have those, you shouldn’t be so concerned about how much money is ready to be deployed in theory.
We’ve even lived through this at least once before. In the FTX era, money was deployed profligately in hopes of leading to good projects and good ideas, which relatively rarely materialized in the wake of that money.
It’s an “if they come, we will build it” mentality, and it runs counter to my basic story of EA’s most compounding and foundational success.
GiveWell started on a comparative shoestring. Just two guys trying to do the rough and ready research to get the best answer they could generate to a question they thought was important: how do we scalably save the most lives for the fewest dollars? Their rough-and-ready answers and the packaging they put around them were good enough that a billionaire couple decided this was the kind of insight they wanted to guide their giving.
Fast forward a little over a decade, and most people in EA’s orbit want assurances that they’ll have high levels of long-term funding before they do their rough and ready MVP or even prioritize building pitches for MVPs.
And look, I know it’s hard to come up with good ideas and even harder to make them concrete enough to be something you could make into a compelling pitch. And it’s harder still to line up the right people with the right backgrounds, raw talent, chemistry, and trustworthiness to turn early funding into scale. I just don’t see having the funding (implicitly) promised up front as making any of that much easier. It probably makes the later stages of this even harder because you get adverse selection for people who see dollar signs.
So I expect most people who try to do this to fail. The low hanging fruit in social good is less picked-over than that in areas where you can generate big profits, but it’s still very hard to find and effort alone is no guarantee. Still, I don’t want people to give up on doing or supporting good things — or being a part of a community that does — just because they haven’t succeeded in what I’ve defined as “the thing” here (seeding new impactful projects through direct work). More successes means more trials and failures too.
My point is that the risk of failure high because for reasons other than funding. There’s $5.2 trillion in savings in the US. All of it is looking for the right consumption or investment opportunity.2 $600 billion of that goes to charity every year and yes, I know, religion and universities etc etc, but you’re telling me you can’t pitch a plan and a team well enough to get a hearing from even 1% of that (in hopes of landing 0.001% of that)? That you’re waiting for some (additional!) billionaire to spontaneously declare sympathy with your perspective first? Or for one of your friends to win the lottery (read: cash in their AI equity)?
The lesson of GiveWell is that great direct work doesn’t draw down some fixed pot of EA money or tradeoff with earning to give; it provides the proof of concept that makes new funders excited to take more risk with EA projects and incubate them for longer. Accordingly, I see the driver of really great projects as having courage in your convictions despite the absence of pre-committed funding.
So in an appalling display of laziness and mild hypocrisy, I’m getting to work on my project of helping people find the courage of those convictions. More to come.
Disclosure: my own idea is mid, but I might have an amazing person to execute on it.
Yes you heard me right, your impactful org doesn’t even have to be a nonprofit