I'm not sure you got my point entirely. If you've got a machine sitting there than can help you, you should use it.

If you're going to write about operating systems, you're an academic fool if you don't also actually take some time to gain some real experience on the very-cheap systems that can help validate that you actually do understand what you're learning about.

If you're going to do AI, you're an academic fool if you don't take the time to implement some of your theories and see how they actually perform. Who cares what someone hypothesizes may work in the field of computer vision if they have no implementation?

And so on and so forth. Even if it is just math, you are an idiot if you pass up the opportunity to implement something to verify your theories.

And the result will be courses, especially at the undergraduate level, that turn out to be useful both for academics, and for programmers who want to learn how to write theoretically grounded code. Which, remarkably, is exactly what we actually have.

You are not benefiting by trying to split academics off from practice. It's a stupid idea and a stupid ideal. It isn't a win for anybody.