That empiricism doesn't apply well to ethics is part of PG's point; he does not believe that the philosophy of ethics is useful. From my definition, I would say that the philosophy of ethics can conceivably be useful, but rarely is in practice.
In vague support of argument one, you might try turning your other argument on its head. Philosophy and science (and, before science, engineering and early health care) developed in parallel over the course of human history; one of them primarily studied abstract problems and tried to make them more useful, one of them studied uses, and worked toward abstraction. Until the 19th century, it appears that abstract study was better for developing new fields, and concrete study was better for development of technology. After the mid-19th century, philosophy lost its firm grip on most forms of abstract study, and became (for the most part) the field that it is today; now other fields tend to be the starting point for the spawning of new fields, and tend to be better for the development of technology.
With regard to your counter-example, I will continue to present my counter-counter-example: mathematics.
Your final sentence is nicely ironic. If, as you've said earlier, you need bad philosophy in order to determine what the good philosophy is, yet PG's work is useless, then PG's work must not be bad philosophy.