When basically everyone in the world faces the same problem, it is interesting if everyone does their own working out to solve it. Even if there isn’t a one-size-fits-all answer, you might think it would be massively more efficient to do a good job of answering once with some free variables.

For instance, my impression is that even before covid, basically everyone was sometimes possibly a bit sick and had to decide whether to go to an event, given that they might be contagious. The usual thing to do, I think, was to consider the question for a bit yourself (or not, depending on conscientiousness), and to maybe mention it to the host and try to guess how annoyed the other guests would be, and then to make a call. But this isn’t a highly personal question—if someone who knew more about infectious disease contagion than most people (which wouldn’t be hard) made a form to tell you what to do based on your symptoms and the number of people at the event, you would probably already be making a better choice, faster, than most people.

Why doesn’t this happen?

Some answers I can think of:

  1. It would be great, but it isn’t anyone’s job, and it’s hard to profit from it because hardly anyone will give you a cent for making their judgment for them.
  2. It would be more effort to open the calculator than the thought most people put into this. (Relatedly, all apparent examples of this are just ‘how worried should I be about X’ where X is something that actually nobody is worried about unless they have an anxiety disorder. There is a form for solving this and it is called a psychiatry intake form.)
  3. If someone made such a thing, there is no way that more than a minuscule fraction of the world would come to know about it.
  4. Such things exist all over the place, and everyone else is using them and have just never mentioned them to me
  5. Such things really do involve a lot of personal information somehow
  6. Few people would trust such a tool enough to do what it said
  7. This sort of thing just isn’t a thing, and people mostly do things that are things. The market isn’t efficient. Everyone’s personal lives are currently like pre-Moneyball baseball.
  8. Nobody knows the answers to these questions. It’s possible to have a better answer than most people’s made up ones (e.g. by averaging most people’s made up answers, or having an expert make something up), but nobody wants to either stand behind that, or pay for it.
  9. This case is a mysterious anomaly, but there aren’t many such cases. For instance, can you think of more good ones right now? No.

Figuring out the covid risk of different activities is another example of massive potential replication of effort, though in that case microcovid.org (and perhaps others) stepped up.

I have a similar feeling at a smaller scale about speculating about the future of AI, which is perhaps why I run aiimpacts.org and other people don’t. (While that enlightens me on some specific things that might go wrong with trying to do this, it’s hard for a particular project to shed light on the problems with a category it is in, since it is in many categories and presumably might have problems from all of them.)

P.S. Yes, I believe the word is ‘octobillion - upling