Often the excuse used to do things properly the hard way even though you're just starting out, is to prepare for unexpected growth.<p>But I like the sentiment with this, and it does seem advantageous to get used to reforging new systems every so often when you reach a certain threshold of demand, rather than spending all your money on building the perfect foundation of a building you might not be able to finish.
Postgres is great but it also seems like a religion at this point. I agree on not using a bunch of software you don't need. But he is saying that it's not simple unless everything is in Postgres. Seems like a premature conclusion.