If you want to attract those kind of people and have them do their best work, then part of the “compelling work situation is to let them pursue their own problems.
Alan Kay:
” I don’t run CDG, I visit it. [Xerox PARC founder Robert] Taylor didn’t want to hire anyone who needed to be managed. That’s not the way it works. I have people on my list who are already moving in great directions, according to their own inner visions. I didn’t have to explain to these people what they would be working on, because they already are. Bret Victor has already hired four people that I didn’t know about. I wanted people to fund, not manage.”
https://www.fastcompany.com/3046437/5-steps-to-recreate-xero...
This should be on the wall of every developer's workplace.
Personally, I like this one.
Ephemera and arcana I've learned decade ago still serve me well today, even when it's outdated people mostly either reinvent stuff so I can recognise new stuff as a variation on an existing thing or build on previous solutions so I know the details and circumstances that lead to some developments and this lets me understand new things better as well (puts it in context).
I think the usual back and forth about developer interviews may have this thought as an underlying assumption; specifically people who can solve a red/black tree on the whiteboard at the drop of a hat must be filled to the brim with useful but potentially obscure knowledge. (Obscure to those who haven't majored in CS or a related field.)
It would be interesting to submit an ask hn "What is your so there I was doing x, y, z when being able to answer obscure interview question seventeen saved the day" story. (That's not well worded but you get the idea).
The majority of programming books are just ephemera and arcana and details that will be irrelevant in a year, or next month when the new version of the framework comes out.
Kay points to books, like the original Lisp Programming Manual, that will help you understand deep core concepts about computing itself, that will remain applicable no matter what framework or library you need to use tomorrow.
Take an Alan Kay, a McCarthy, Norvig, Abelson, Sussman, Armstrong, Steele, etc. from their prime and drop them into a software company where they have zero familiarity with the programming languages or tools currently being used, and within days or weeks they will be the most productive developer at that company by far. They will come up with simple, elegant, high performance and correct solutions to problems none of the other developers would have even considered.
Those are the kinds of thinkers you want to emulate, if you really want to write excellent software solving real problems in the shortest amount of time.