I can easily see why the former is true. The latter seems a lot less likely.
I can easily see why the former is true. The latter seems a lot less likely.
The author cites examples such as Linus and Margaret, but IIRC they studied CS and/or math as part of their educational upbringing... so I feel like they're almost counter examples of what the author is arguing for.
It seems like the author is really championing the "self-tinkering engineer" as the outperforming engineer.
What is studying if not self-teaching? But, semantics aside, Linus, at least, was programming long before he attended university, and his claim to fame was built early in his time at university. It is true that he "studied CS", but that came later. He is unquestionably self-taught in the area for which he is best known.
Those self-taught engineers who don't even perform aren't going to be engineers for long. So of course you'll see a lot of self-taught engineers in the outperforming category, it's necessary for survival.
As a self-taught person on a lot of different matters, I find myself exploring rabbit holes that expand my knowledge, but don't progress the task I originally started doing.
To your point though, I think it doesn't matter so long as you've learned to deliver business value. Application of broad and diverse skills may deliver value at a start-up for example, but wouldn't get too far at a ticket shop.
However, someone who already has the talent to be really good at something and who has the inner drive and motivation to push themselves is someone who is likely to excel. So if you find someone who is excellent at something and self taught, it's not a surprise. They probably combined natural talent with a strong work ethic, and lots of exploration of the entire search space.
Also, this (and other things I’ve read) always seems to argue against the strawman that “you need a formal education to do well in software “. I’ve never seen anyone say that (including during many years involved in hiring at big tech). The argument is that the pool of CS graduates are more likely to do well (and bigger and simpler to find) , so it makes sense to focus there if you need to hire a lot of people.