The real odd thing here is that `work_mem` (and to a lesser degree `maintenance_work_mem`) is still low enough out of the box that a script got written to highlight that it's too low. Increasing it is the first step one probably ought to look into when installing Postgres.
Huge numbers of postgres installations are on small machines and/or machines shared with other software. And whether you want a large work_mem hugely depends on whether you want to run an OLTP workload with a few hundred connections, or an analytics workload with like 2. It ain't as simple as you paint it.
