Light’s still slower in glass than in a vacuum, though, or it wouldn’t refract at the transition between the two.

Right, and it varies with what type of glass, google "refractive index of glass".

If you google image search "refraction fiber optics" you'll get some decent pictorial examples of what I meant by photons bouncing off the interior walls of a 9/125 SM fiber optic strand billions/trillions of times on its path, I guess I was trying to write an extremely simplified explanation of refraction in something like a typical SMF-28e / G.652.D fiber.