There are serious disanalogies, too, but I still like this, from Eliezer Yudkowsky:
Lots of superforecasters [link] are programmers, it turns out, presumably for the same reason lots of programmers are correct contrarians of any other stripe. (My hypothesis is a mix of a lawful thinking gear, real intellectual difficulty of daily practice, and the fact that the practice of debugging is the only profession that has a fast loop for hypothesis formulation, testing, and admission of error. Programming is vastly more scientific than academic science.)
And later:
You’d need to have lived the lives of Newton, Lavoisier, Einstein, Fermi, and Kahneman all put together to be proven wrong about as many facts as a programmer unlearns in one year of debugging, though admittedly they’d be deeper and more emotionally significant facts.
I think there’s something to the spirit of this quote, but one serious disanalogy sticks out: programming is vastly more “scientific” than academic science because it deals only with an incredibly simple, lawful, constructed world. I reckon a number of false contrarianisms common among programmer types have their roots in this (e.g. some forms of libertarianism).
It’s a little like saying that Zendo players are the world’s greatest scientists; they’re only *that particular Zendo instance’s* greatest scientists. The overall “attitude of science” might well be shared in common, but the scale and thorniness and slipperiness of the problems are so far apart that the similarities are the exception.