[https://www.henrikkarlsson.xyz/p/ai-tutors]

In other words: the main bottleneck in education, now that tutoring is becoming cheap, is culture design.

Can we figure out ways to scale access to high-growth cultures? Are there ways for more people to grow up in cultures that approximate in richness that J.S. Mill, Pascal, and Bertrand Russel had access to?

It is not a question that has been given serious consideration in the pedagogical literature that influences the design of schools. But there are insights to be mined from anthropology, internet community moderation, lineage traditions in martial arts, small experimental schools, scenes, workplaces, and so on. It is possible to deliberately create high growth cultures, and software makes it easier to scale them. But there are many open questions. (If there is interest, I could do a write-up of my understanding of this design problem.)

Cultures and tutoring systems are complements. The more powerful AI tutors become, the more valuable cultures that support learning will be. (Teaching and tutoring systems, on the other hand, are substitutes, which implies that we can expect teaching to lose value over the coming decade.) There is an opportunity here to direct resources away from things that AI systems can automate, such as teaching, and into the more ambitious project of building better cultural infrastructure.

To be clear: I’m not saying that AI tutors, or other kinds of software, can replace humans. Most of us need the support and community of others to push ourselves to excellence and to find meaning in the projects we pursue. But soon we will be able to spend less precious human time on basic tutoring; instead, the emotional labor we do to support each other can be invested in a more leveraged way.

The education system will be slow to embrace this possibility. But others will fill. We will tend the bull.