Once you have an app, a lot of data, and actual problems, it's far easier to pick the right alternative.

PostgreSQL is good enough to get to medium sized with nearly every use case. Once you are there, you have the use case and the test data to test any alternative for it well, rather than trying to guess beforehand what you actually need.

The advice is basically "PostgreSQL is probably good enough for whatever you're building now, and you should only look for other solution once you are big enough that it stops being that"