Some of Ord’s estimates have data backing them up — we know a bit, for instance, about past supervolcanic eruptions and asteroid impacts — but the threats we’re told to be most frightened of are also the ones that are nearest to pure speculation. Which is not to say the speculation is wrong; I think synthetic bioweapons are terrifying, and the risks of artificial intelligence are worth seriously considering. But I think the false precision can make the unreal feel real and hide the shakiness of all that came before the quantification.
This, to me, is the clearest link between Bankman-Fried’s fall and certain elements of effective altruist culture: Crypto is built on attaching values and probabilities to notional assets and currencies. What looks like a balance sheet from one perspective proves to be nothing but a set of arguments, assertions and thought experiments from another.
This problem exists elsewhere in capitalism, but it is concentrated into crystalline form in crypto markets. Too often, the only real value of crypto assets is self-referential: The asset is valued because it is valued. There is nothing behind it except other people’s willingness to believe the story you are telling them. There is no army or auto factory or even beloved work of art. There’s just code and quantification. The numbers are lending undeserved solidity to an abstraction. I think effective altruists have a tendency to bewitch themselves in much the same way.
We Can Do More Good, and We Should
I worry that this column will be taken as a reason to dismiss any concerns that sound weird the first time they’re heard. That’s not what I’m saying.
Effective altruists have fought hard to persuade people to worry more about artificial intelligence, and they have been right to do so. I don’t think you can look at the remarkable performance of the latest artificial intelligence models — Meta created Cicero, an A.I. system that can manipulate and deceive humans to achieve other goals, and OpenAI’s latest bot will advise you on how to create nuclear weapons if you ask cleverly enough — and not think it important to consider the consequences of vastly more powerful A.I. systems.
But I think too much of the energy and talent in effective altruism is flowing away from the compassionate rigor that initially distinguished it, and that the world still needs. Effective altruism will not be nearly as, well, effective if it loses touch with its early focus on improving the lives of people living today.
All of this brings me back to GiveWell. GiveWell was founded to assess and even produce that evidence, and it does an excellent job. Its research is comprehensive, thoughtful and, most important, transparent. I don’t agree with every decision it makes — I think it sets the bar for cost-effectiveness too high, and some charities that have dropped off its list, like GiveDirectly, remain on mine — but the persnicketiness is the point. I give to GiveWell’s charities every year, and while that’s not the whole of my giving, that’s the part I feel most confident about. Giving to organizations I’m so certain of is a good feeling, and I hope you get to feel it, too.
The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: letters@nytimes.com.
Follow The New York Times Opinion section on Facebook, Twitter (@NYTopinion) and Instagram.