[jholman]

As an algorithmist, I strongly agree. Algorithms are of value, but they are not a high priority for junior devs. For a tiny fraction of problems, algorithms knowledge is absolutely essential, and often cannot be picked up at the time of need. But for almost all problems, especially the problems facing junior-to-intermediate devs, there is rarely any algorithms work. As you said, code organization. Long lists of things to remember to take into account (e.g. what areas require one to stop and get a security expert). Etc.

I think the best thing you can say for algorithms training is that, without it, you often don't notice that you need to get an algorithmist.

When people say, as my sibling commenters do, "find out why your n^3 approach is slow", or "do a substring match", I wonder what problems they're solving. Has this shit actually come up in your career, or is it just something that you imagine will come up? How many times has a cubic algorithm actually been the problem? As compared to doing five hundred DB queries, instead of one query that JOINs properly and returns a 500-row result set, which I see juniors and students do all the time, and is not an algorithms problem at all (it's the same asymptotic complexity, btw!). How many times have you tried to figure out a fast approach to substrings, and how rude was your code reviewer when they pointed you at the standard library?