I’m not the first to say this, but I’ll say it again: I’m worried about Anthropic’s and OpenAI’s (possible) IPOs.

Tl;dr: there’s a good chance that Anthropic and/or OpenAI becomes a publicly traded company in the near future. Somehow this makes a lot of capital that was previously solid melt into a liquid. That liquid capital can then be given to whatever the people who own it want it to be given to. That’s many billions of dollars in annual funding at the very least, which is orders of magnitude more than organizations are used to receiving. I’m thankful many of these people want to give their money away, though really that should be the bare fucking minimum for billionaires. But I’m also worried.

We are all attracted to money like moths to a flame. Even I feel the temptation: with all these billions of dollars, what kind of good could we do? We spin great yarns about the kinds of impact we will have on the world and exchange congratulations on our ratiocinations. Yet the more I’ve thought about this, and the more I’ve read what people have had to say, the more worried I’ve become. To say that the current vibe is cavalier is a massive understatement. One of the first-place essays in the Manifund essay competition basically says: come to where all the money is; we don’t know what the fuck we’re gonna do with it, but we’ll figure something out. Come up with a halfway decent idea, and we’ll give you 6-7 figures per year and see if the “long shot” pans out.

It fucking enrages me when I think about it.

There are people in this world who have gone their whole lives without the tiniest fraction of what these Silicon Valley types let slip between their fingers. People who can’t buy a cart to hold their fucking vegetables. People who sell their children into slavery because starvation is the alternative. People whose children are dying from malaria because they can’t even get nets to protect themselves. And these “philanthropists,” so-called lovers of humanity, think that a hare-brained scheme deserves that money more? It’s fucking sickening.

GiveDirectly should always be the reference point here, because it’s literally as simple as giving money to people who actually need it. If your program can’t come close to saving a life for $10,000 or the equivalent, then stay home and do your fucking job. OR DO FUCKING BETTER.

We’ve seen how this kind of thing can fail before. FTX threw tens of millions of dollars into stupid shit like the halftime superbowl ad and various celebrity endorsements. Even many of the less stupid ideas, like the Carrick Flynn campaign, were costly and ultimately fruitless embarrassments. And then, of course, FTX and Alameda Research got exposed for fraud, and the EA movement did a surprised pikachu face. All of that promised windfall just blew away. And all of us were left to pick up the pieces. Though before we pity ourselves too much, let’s remember that we probably weren’t the most viscerally affected.

It’s not just EA: it’s a broader cultural and institutional problem. I was outraged when I learned that in the wake of the 2008 financial crisis, no individuals were ever held accountable. The banks paid ever fatter bonuses to their executives, none of whom were even charged AT ALL for their role in the crisis. Not even a slap on the wrist. Meanwhile, ordinary people had to suffer the foreclosure of their homes and the loss of their jobs. But that’s just what happens in a society of limited liability but unlimited profit.

Nassim Nicholas Taleb was right to call us back to the ancient principle: the man who builds a bridge should be comfortable sleeping under it. It’s for the worse that our society has so enabled “fragilistas” (myself included) to set policies without cowering under the sharp blade of genuine consequences. Maybe there’s a small chance that society figures out the fragilista problem before Anthropic/OpenAI does its IPO thing, but in the event that it doesn’t, we’ll have to entrust the distribution of those billions of dollars to a few elite fragilistas who have no skin in the game; whose median worst case scenario is probably having to move back home with their affluent parents. To put it lightly, this is an incentive structure that does not inspire trust, and it is thus up to individuals to create something worthy of trust despite that. So it is to these people whom I now speak.

To the people pitching ideas for this windfall: think of the dozens, if not hundreds, of lives your ideas will cost in opportunity. And then think about whether it’s still worth it, given that.

And to the big funders themselves: I hope you’re not fucking sleeping at night. You know as well as I do that people will worship you if this windfall is distributed anywhere near successfully. So don’t worry about that. Worry about whether your donations, in total, will do as much good as saving at least 30 million human lives.

That’s right. 30 million lives.

How did I get that number? Well, a recent estimate of the size of the windfall puts it at about $370 billion in total. An estimate of GiveDirectly’s effectiveness puts it at about 7.7 WELLBYs for $1,000, or about one life for every $10,000. You can nitpick the math, but the point still stands.

You have all the power here, and therefore, all the responsibility. This isn’t some kind of game where you exchange your casino chips for social credit points. You are playing with peoples’ lives.

And if you mess it up, know that everything that comes after will be on your head.