[nowherebeen]

I think yes and no. After working for 7 years+, you will hit a plateau. To break through, you will need to study on your own free time. That's because most individual work doesn't get harder after a certain level. It is the scale of it that makes the work difficult.

[capitalsigma]

Not disagreeing that self study is valuable, but it's unclear to me how writing a bitonic sorting network in a bespoke assembler is going to get me promoted to staff

[agentultra]

It generally won't.

Sort of like how lifting weights won't train you to be a better line backer on an American football team.

You lift weights because it strengthens the body and you need a strong body if you want to have a long career as a reputable line backer.

(Maybe a bad example... head injuries and all)

Knowing how to implement combinatorial algorithms or SAT won't get you a promotion unless your job is implementing libraries of code that use those algorithms... but it's hard to see the forest for the trees if you can't recognize them... so to speak (Knuth loves trees). It can level up your mind so that you can solve more challenging problems or find innovative solutions.

I don't really read through AoCP with a job promotion in mind. I mainly do it because I find the subject matter enjoyable. And I only work on parts I find interesting or when I come across something I've heard about before but don't know well: chances are there's a data structure or algorithm explained in detail in AoCP.