To me it makes more sense to call something that doesn't need a separate server "serverless". It's quite weird that the usual use of "serverless" means something on a server that you don't manage.

Ha! I agree! That usage seems to presume YOU are "serverless" not the application or the system running.

In the context of (cloud-native?) databases, serverless seems to really mean something else entirely: https://me.0xffff.me/build-database-in-2020s.html

But to me, serverless the way sqlite is, also makes ample sense.

And yet, at the same time, it does make some sense, as they took the “server” component away and packaged it as a regular cli tool.

I’d argue that it’s actually more accurate a description than the more common use of the serverless definition.