You are minimizing the criticism, which isn't just about some utterance at a convention, it's about the core premise on which Chomsky's entire work rests, namely determinism. And more generally, that success of a model should be tested empirically rather than that it should "provide insight" in some abstract way. There is no doubt that Chomsky has done great work in the area of formal languages, it's just that the relationship with natural language is not as strong as initially hoped.