But it's not like he wrote RethinkDB and then wrote Postgres with the benefit of hindsight. When he says that the big change was the database and not his own code, I believe him.

I believe him, too. I'm not trying to participate in any database war. I've only said that any rewrite is going to have the benefit of hindsight, and because of that, his results are going to reflect that to a degree.

He outlines several places where he had to do something because postgreSQL doesn't do it, but he suspected before hand that RethinkDB didn't do that well.,,

The article talks about more than mere code efficency.

I don't mean any more or less than what I said. Rewrites are going to have the benefit of hindsight.