Self-awareness is always a good thing, no?

Kuhn's key point is that scientists (like many humans) tend to be herd animals. When you get results that are inconsistent with the public record, not to mention current dogma, there's a tendency to dig seriously for systematic errors. And when you start getting results that are consistent with the public record, you tend to get on with the writeup. Being aware of that stereotypic behavior is the first step in being free of it. No?

I agree, but doesn't this mean that scientists should read about the history of science, or sociology of science, not philosophy of science?

All of them!

And anyway, those are all artificial categories, I think.

See also Barnett Newman's use of the same metaphor in reference to aesthetics.

I attempted to address this above, in a reply to a sibling of your request.