The first one that comes to mind is something I haven’t done lately but used to do constantly, enough so that I wrote a custom fuse filesystem to combine isofuse with http handling, and that’s parsing an index from something with a range request then fetching only the part or parts I want. Think fetching only the trailing index from a zip archive and then fetching the specific file you want out of a 10tb zip rather than downloading the whole thing. Now, how often would I do this on the command line? Not sure, but I might not have bothered with fuse if I’d had the option.

Pausing and resuming downloads, and fetching bytes in splits would turn curl into a download manager like aria2.