In Greek you can use the same root for different purposes. I.e. it is not the word sophia or sophos that came to denote deception or fallacy. Sophist is someone who deceives using the pretense of philosophy.
I do not see knowledge as something that comes with an application manual. I.e. there is no manual defining how will use number theory, but still it has applications, sometimes completely unexpected.
I would argue that philosophy navigates the fields where the testability of equations and "hard" science fails. That is also why some scientists do philosophy when/where the rules break down.
I would agree that philosophy has evolved somewhat (Nozick and Rawls), but for me it never really achieved the level of empirical usefulness (past perhaps rule utilitarianism), because it has never been clear enough how (or even when it was appropriate) to apply it. If there were such a thing as "Newtonian" philosophy that had useful testable equations, perhaps I would view it differently.
At least I have some idea when physics rules break down and when they are useful approximations. In my experience most philosophy is written as if it is completely true not a model of reality.