i don't usually talk about tech gripes, but i think occasionally about the time somebody told me a few years ago it would take upwards of five hours for a job to process a million database entities and i responded, without thinking, "on what, a pentium?"

the answer is that you can do this in a reasonable timeframe if you don't ask your extremely tired primary db/server cluster to do this massive data ingestion and processing work and can then gradually fold normalized data back into your cluster in a bunch of more trivial operations, but i still think that "is this on par with what i'd expect to see on a Pentium data sheet" can reveal unexplored perf or platform arch issues.

a big refrain from me is often "machines should work; people should think" which is shorthand for "be pragmatic, the computer can do it fast enough" most of the time but also "if the computer is being comically slow, it's because we haven't thought through what's happening" the two can exist side by side in any case.