Their loss I suppose.
Our loss, really: less good quality games for everyone.
I threw in the towel at GameDev for reasons very similar to yours ("premature optimization" also comes up far too frequently) - from, what can only be, people who have never even worked on specific components of games that they are claiming knowledge on.
The StackOverflow system somewhat fails on GameDev, because there aren't enough knowledgeable people on the site to issue corrective votes (likely because these people are scarce in general).
Incidentally the bad advice is bad on many levels: even if you somehow ignore UDP it's still bad advice to do "RPC over TCP". TCP works best (fastest, least latency) when the send buffers are full at both ends - which RPC, by definition, does not do. This means that these people simply couldn't have researched the fundamentals of the protocol that they assume to be "good enough," how could they therefore be aware of how it is better or worse than alternatives?