No one is assuming anything about his OO skills, they are reading the words written and assuming the author isn't lying.

Several people in HN should know more about OO than Armstrong.

There are academics that have taught and worked for decades on OO systems and concepts among the crowd, unlike Joe.

In general, popular language creators are hardly the more knowledgable of the various PL concepts and issues. Simon Peyton Jones probably knows hundred times more about such stuff than Matz or Guido do, for example.

I do not "insist" that he is a "dilettante", I question people on this thread quoting him as an expert on OO as there without showing that he is an expert on OO. Beside I give circumstantial evidence why he isn't ("Tried Java ...", "Threw C++ book at the wall ..."). I think it's important to differentiate strong opinion from expert.

And there are more indications: His understanding of "isolation" in the quoted text above is very influenced by him being the Erlang creator and the way Erlang thinks of isolation (processes). None of the (C++/Java/Ruby) OO programmers I know would use "isolation" in an OO context that sense (one error crashing other programs in the same VM).