The wind sprint, the grade as a breaking lever, and why the conditions for struggle turn out to be relational, not cognitive.
Most conversations about AI in schools begin with cheating. It's the obvious worry, and the wrong starting point. The better question is about mental fitness: what happens to learning when the struggle that builds it becomes optional?
Teaching has always run on proxies — the essay, the problem set, the confident answer in seminar, all standing in for learning we can't observe directly. AI loosens the link between the product and the mind that made it. That doesn't lower the stakes so much as raise them: it asks us to teach with more clarity about what we actually want students to become, and where the effort has to land.
I work on that from two directions — writing about it, and building the tools that test whether the thinking holds. Neither is solved. Both are worth the attention.