My gripe is that the Finder is still full of O[n] or even n^2 algorithms, which worked fine in 1990 when people had maybe 1000 files on the entire filesystem. But nowadays, 10 million files is a small file system. And O[n+] just doesn't cut it.

Sometimes you can't do better than O[n], in which case you sweat the details of that inner loop very hard. Finder doesn't sweat much of anything.

Any time I have to do something to a few thousand files or more I switch to the Terminal and use the command line because Unix [mostly] knows how to deal with large numbers of files efficiently.