In fact, Unger is mixing up various positions of Wittgenstein over several decades. Usually, Wittgenstein is treated as two authors, numbered Wittgenstein I (the author of the tractatus logico-philosophicus) and Wittgenstein II (the author of the Investigations and beyond). Now, the position of rather teaching, was Wittgenstein's position after he thought of having ruled out any questions that would matter to be answered by philosophy (in the tractatus). The Investigations are quite opposed to the project of the tractatus, starting to investigate the production of meaning by language as a social process, dismissing most of the notions of the tractatus. (In other words: While the tractatus was using meaning for asking, if any traces of "the mystic", the spiritual meaning of the world, were to be derived by logic – short answer: No –, the Investigations have an all together other topic, how meaning itself would be derived, contesting the very basis of the tractatus.)

And on metaphysics in general: We have learned so much more from Kant than, say, Da Vinci; what would Unger have to say on engineering? If philosophy isn't about easy answers, it isn't about easy questions, too ...

P.S.: I may be biased, since I've a degree in the field. Yes, there's much crappy philosophy, esp. in the 19th century. But then, there's much crappy software, too. Should we give up programming, just because MS failed again?