>> can't remember a single class where a professor actually walked us through the process of figuring something out
> This can't really be blamed on the tool of the chalkboard.
That is true, of course, but that was not the point. The point is that blackboards do not necessarily produce, or even encourage, good pedagogy. The two things are largely orthogonal.
> Writing code is a narrow and constrained tool with much less freedom than writing on paper
Yes, I think this is a feature, not a bug. It enforces a certain amount of rigor that blackboards don't.
Of course, the language in which the code is written matters a lot.
> can't remember a single class where a professor actually walked us through the process of figuring something out
This can't really be blamed on the tool of the chalkboard. I promise you this does happen in some classrooms. It certainly happens all the time in informal discussion between professionals. In my opinion schools should strive much harder to induct students into a problem-solving culture than teach specific content trivia. A blackboard is an invaluable tool for the former.
> often solved by writing code
Writing code is a narrow and constrained tool with much less freedom than writing on paper, and thoughts made first in code are substantially blinkered in my experience compared to thoughts worked through on paper. Many of the best programmers I know do their most important work on paper (sketching, planning, doodling, drawing diagrams, ...) before ever touching the keyboard. And that is far more true still for mathematicians, physicists, etc.
What is true is that many programmers are not very effectively taught to use paper to its full advantage as a tool for thought, and either never learn or largely figure it out on their own. Our typical computer science pedagogy is frankly terrible at this, and professionals' on-the-job training isn't much better.