I miss when posts like this mattered.

That's not to say performance doesn't matter anymore or that blog posts on niche topics don't matter anymore.

It's more that there are 30 opponents on all sides fighting to minimize the impact of this kind of post. CPUs are still getting faster even now despite Moore's law being dead. The business or career impact of choosing between an associative list vs hashmap in a garbage-collected language like Guile Scheme is so minimal that it's hard to quantify.

If it's in a hot enough path that it matters, it's likely that there are at least 3 things you can do within 20 minutes of work (or 5 minutes of GPU time) that will solve the problem as effectively or better.

I remember the very specific period of time when blog posts talking about functional programming for React developers were en vogue. You can speed up you Scheme app by 15%, or you can build and deploy a completely new service from scratch in Node.JS in the same amount of time.

It used to feel like code had some kind of meaning or value. Now, it's just an artifact produced as a side effect of work. But that's been a trend for a decade or so now, AI is just the latest (and most significant) iteration of it.

Sounds like the difference between code as a craft versus artifact and product. Actually not even product, it's the inner guts of a product that most people don't need to care about, and increasingly just machine generated so it's not even meant to be read by humans. Write-only code of the post-programming era.

Professional software has always aspired to be an industrial process, like OOP and Agile, as a collective endeavor to produce code of decent quality that works reliably, to achieve business goals. Any aesthetic satisfaction or philosophical insights are a byproduct, nice to have, but not the main point.

Code as a craft is a niche for experts and researchers, for hobbyists and amateurs. The miniscule performance improvement gained from choosing an array or hashmap is insignificant in most situations, other than maybe resource-constrained contexts like embedded programming, retro computers, games, competitions.

But, thinking over it, code as a craft still has a healthy subculture of people across older and younger generations. Perhaps it's up to the older ones who remember the good ol' days of respectable craftsmanship ("craftspersonship") to promote and encourage others to carry on the tradition.

Or is that not even worth doing anymore with language models pumping out vibe-coded slop? Will programmers be relegated to reviewing and fixing such mass-produced code? Yes, probably. But no, of course it's worth preserving and evolving the culture of computer science and software development, maybe it's more important than ever to keep the flame of human spirit and imagination alive, supported by machine intelligence rather than replaced.