It is unfair only in that it leaves off an important qualifier: he thought it was pointless as far as his theories went. That is he thought that there was universal grammar (actually not quite a grammar--but lets stick with his earlier oversimplification) and that would be the same among all peoples. He never said it would not detect things of interest, just not things of interest about HIS theories. (Of course he also wanted to define stuff that went much beyond his theories as not being linguistics, so perhaps not so unfair.) However of course it seems that these empirical tests have revealed something of interest about his theories: that some of them may not be true. Even if on further examination they end up not disproving any of his later theories, it was perhaps a bit arrogant to assume this was beyond the realm of possibility. None of this changes the fact that some of his basic theories remain the foundation of modern linguistics.
-- Please note: Personal messages should be sent to [garlpublic] followed by the [at] sign with isp of [comcast], then [dot] and then an extension of net