I always thought that the peter principle is very flawed, in that it assumes an increasing level of competence required per job, and people whose qualities are static.

About the first, being manager is not inherently a more difficult job than being a senior IC. It's just a wildly different skillset.

About the second, reaching a position you can't do does not at all mean you'll never be able to do it. We learn.

The right insight there is that the switch to management as a necessary step to grow is horribly handled. But "leader as a profession" sucks as well, managers fundamentally ignorant of the kind of work they're managing.

I think what we need is a more fluid relationship between work and learning. We assume a college->work pipeline, when maybe we should have career stages at which we go back to further learning.

All of this is perfectly covered by the Peter principle. It doesn't assume an increase in difficulty at all, it just assumes a different skillset. People are promoted (maybe to manager) based on their performance in their previous job (maybe programmer). If they can adapt, great, and maybe they are even promoted further. If they cannot, they have reached their level of incompetence.

Why would going back to college help here? You don't need more certification you're already in and proved your value. As an outside and self taught I've always been fascinated. For a subset of people is it actually learn better in college style environments than on their own?

In most (all?) jobs, you can’t make more $ in the same role outside of inflation adjustments. You want more, you need to move up or hop. Moving up separates one from what they were good at, and enjoyed. All of a sudden you are dealing with vacation requests and some HR portal.