> You can drill into kids

The thing is that one cannot. Or at least there are no idications that those drills have any significant effect on a later life fading away pretty quickly a year or two after school/college. At best during those drills children learn from teacher behaviour, not words. I.e. they learn how to force others to do things they do not understand or need.

Again, how are you measuring this correlation? If it's not something that's measured, you don't have any numbers to correlate with.

You could equivalently say that scores in any particular subject in school don't correlate with success after school, which is usually true. But it doesn't follow that net success wouldn't change if that subject was taught better: not only would you be teaching different techniques than is measured in that correlation, but you also have network effects when more people are trained in something.

An easy example of this is the topic at hand: software development as part of a team, which many schools don't teach (or don't teach well), but is a skill picked up quickly (hopefully) when entering a job with more than a handful of developers. I'm not talking pair programming or anything specific, just generally being able to develop software with other people.

Here you have something that isn't measured at many schools and is important to success at many companies. So you'd find no correlation between the two, but better developed experience would undoubtedly help the worker and the team in that situation (whether or not that's a skill best learned in school vs on the job is a different question).

> At best during those drills children learn from teacher behaviour, not words

Your apparently data-based argument is wandering into gut feelings, but this is what I was talking about. If a teacher lionizes Albert Einstein for single-handedly saving humanity from 19th century luminiferous aethers and inventing the atomic bomb, the idea that progress is only made by hermit inventors is going to continue to propagate. People are born to be math geniuses, they don't have to work at it, etc.

As a part of development team one learns precisely from behaviour of others when one is not a subject of a learning drill. The book in [2] from my comment above was rather convincing that this the most effective way to learn. And such learning does not imply that one learns team work, as one could learn habits of a single most harizmatic person, not from the whole team.

As for teachers lionizing something, then consider that in Soviet Union in school and university one was constantly subject to drills about Communist Party. It did not help the idea to survive at all.