It predicts how much time they spent solving leetcode, and biases towards recent graduates. If you’re hiring senior talent it’s mostly a wash and a waste of time. But it’s deep in valley culture so we’ll likely waste our time with it for another decade or so before we wise up.
For myself at least: absolutely yes. That's actually some of the first stuff they taught us. Would I want to be interviewed on whether I can implement a linked list from the top of my head in an arbitrary language and 100% correctly in a high stress interview situation? Absolutely not. (and a linked list in the end is something really really easy if you ask me)
Does that in any way relate to what I'm doing right now? A little! This will differ depending on who exactly you work for and in what role. It certainly helps to have learned this stuff. I can honestly say that it was fun seeing how regular expressions actually work, how you implement a parser for it etc. But then you learn as well (right there in university) that you better use a parser generator after just having defined the grammar instead of writing it yourself. How many jobs are out there that require you to write a grammar for something? To create a new language? Very very few. There's a market for "boring business software X" software developers.
Is it the most important thing I consider when hiring? Absolutely not. We also don't pay FAANG salaries where I work. Do I want capable software developers? Smart guys for sure. People that are curious, willing and able to learn, writing maintainable code. Guys that if I tell them that there's a hidden n^3 algorithm in what they just wrote and it _does_ matter for the (relatively small but large enough) volumes of data this will run on can look up/ask for what that means.
EDIT: interesting down votes. I hope someone will reply to explain. I certainly don't see why we can't have a civil discussion especially given that we seem to agree on it having been good to get a solid university education. FWIW I have a Master's in Computer Science. From a country that values non university education (Germany - where you can do vocational training to become a programmer. Just like you can become a master carpenter through that system).
It’s also really downplaying how much one can learn when they need it for an application - yes it’s true you have to know about something to go after learning it, but there’s nothing stopping someone picking up a book when they get stuck to learn about what they don’t yet know.
Much of a CS course is nice to have knowledge for many programmers, and can be plenty beneficial. But a CS degree is there to show you theory not teach you how to program - the amount of people with degrees that are very poor at even basic programming is not as low as you’d think.
I work on a very large system, but spend far more time chasing down poor performance due to emergent behavior than I do worrying about Big O on some small piece.
Many of the best programmers I've worked with haven't.
Would I be a good CS researcher? No, not even close.
Am I a good software engineer? Damn straight I am.
Would I work with you? Sounds like we wouldn't be a good "culture fit."
I don't like it, but it's just the way it is now.
I imagine we operate in very different geographic regions of the USA.