He suggests students can be more intuitive about general training than economists sometimes. It's interesting, but I'd suggest the story runs a little differently. The student's answer, of course, is a classic answer by economists to the problem. So what's Bryan's concern? The concern is that we all still cite Becker, essentially.
But that's because of the way we structure these discussions. Normal science is puzzle-solving, so we have a tendency to pose ideas as puzzles and then pose solutions. So the puzzle is that that in a simple world firms shouldn't provide general training but in the real world it seems like they do. There are a number of reasons for this - one is delaying wage increases to split the cost of training (Bryan's answer), or we could also think of bundling specific and general human capital.
Those are classic economist answers, we just approach the question with this puzzle/solution framing. We don't cite Becker because we think that's the end of the story. We cite Becker because it highlights the significance of the response.
This has applications to Fama/Shiller too. We don't cite any of this perfect rationality stuff because it's literally right. We cite it because it's a very reasonable baseline that poses a puzzle and focuses us on what appropriate solutions to that puzzle are.