Good read. Both your and Andy's articles remind me of why Socrates (allegedly) derided books as a lower form of education. When we lose the ability to question literature, we lose the capacity to build knowledge models. In compulsory schooling, this turns knowledge learning into rote memorization.

Books turns the labour of both questioning and answering to the reader themselves. Perhaps it's why I loved reading physics over any other subject. All learned concepts can be tested, even as a thought experiment (I couldn't be bothered to memorize the colours of chemical elements, compounds and flame). My first advise to classmates asking me how to study was always to "build a model in your head".

I would have thought it's obvious to anyone who has studied any subject in depth that you need to actively engage with the material. Books are great because they go at the pace that suits your learning and are easy to skip around. Video not so much.

The proposed replacements seem to be technology that takes some of the active role away from the student. That doesn't seem to me that it would be effective, as anything I've learned well has required I do that kind of active introspection. Off-loading it to a machine feels like it would harm my learning.

Also, very typical of software people to reach for software solutions. There is a much wider space that can be explored. Maybe that would be more fruitful?

I think that Any Matuschak's article is much better than the posted one.. My main problem with this article is that it frames constructionism as _the fix_.I think constructionism has its own problems. I listened to the Chalk & Talk episode with John Sweller recently[1] and he had quite salient criticisms of some of the tenets of pedagogical constructionism.

[1] https://www.youtube.com/watch?v=WPm3ls6y-GE

This is well written. And it doesn't even focus on AI, which is rare these days. However, I don't fully agree that AI and the interaction it brings are just as flawed. I think used correctly, it will be a big help. Of course, like you mentioned it misses the social and emotional connection, but so do textbooks, and the interaction component is much better. To use your own words, it can lower the burden of metacognition significantly. It can handle the learning path and what exercise to handle when. Maybe even informed upfront by the author. Love it or hate it, I think there is a lot to gain in terms of human learning support in the field of LLMs.

> This article is my response to Andy Matuschak’s article Why books don’t work (https://andymatuschak.org/books/). If you haven’t please read it first. However, if you want you can read this article now. I believe the following content should still make sense.