I can believe that as regards science practice, and think a lot of epistemology of science and such isn't that relevant or useful to the practice of science (everyone seems to have kind of a half-worked-out theory of verificationism or falsificationism or something similar that they claim to subscribe to, but things mostly work pragmatically in practice).

I'm more thinking of physicists who, especially in their later career, decide to write what are essentially philosophy works, only they: 1) don't acknowledge them as such; and 2) they are therefore often not as good as they could be, and sometimes bad. For example, Roger Penrose's stuff about physics and consciousness is a bit strange, and it probably wouldn't hurt if he cared more about existing work in philosophy of mind. And Hawking's most recent book has good science, and an interesting philosophical proposal of "model-dependent realism", but strangely refuses to consider it to be a philosophy proposal or discuss relevant existing positions.

Physicists veering off towards "big-picture" philosophy-ish stuff later in their careers isn't new, but this sort of seat-of-your-pants approach seems newer. Looking at the philosophy-oriented works Max Planck wrote later in his career, for example, he's impressively well-read in both physics and philosophy.

I do not think that's true. Are you familiar with the idea that scientific theories have to be falsifiable? That was formulated by Karl Popper in the mid-20th century.

If anything, the rise of analytic philosophy and the introduction of greater logical rigor to philosophy in the early 20th century made it relevant again after a largely fruitless 19th century. Go ahead and ask a scientist whether Popper or Hegel is more useful.

It's not like before Popper scientists didn't know they had to design experiments to test their theories. Popper himself, from what I read of him, considered his work more useful as a way to tell apart real science from fake science (he gives Marxism as an example of the latter), than as anything that could help real scientists do their jobs.

In chapter 7 of "Dreams of a Final Theory", Steven Weinberg argues that philosophy has been mostly useless or even harmful for physicists, and whatever positive effects some philosophical theories might have had, had to do with undoing the harm done by other philosophical theories.

Weinberg is not hostile towards philosophy, he has warm words for it and says he enjoys reading certain philosophers, he just acknowledges that philosophy is not at all helpful in doing science. Exact quote: "I know of no one who has participated actively in the advance of physics in the postwar period whose research has been significantly helped by the work of philosophers." (Weinberg considers this surprising, and contrasts it with mathematics, which is extremely useful even though there is no reason why it should be.)

I'll agree that philosophy doesn't really help one do science, but it does justify why it's worthwhile to do science, and real science at that.