Nehaveigur

I’m increasingly encountering the belief that capable leaders have become rarer or have even disappeared. While most often observed in the political realm, some argue that there are also fewer scientific leaders. Erik Hoel makes the point that today there are fewer geniuses of any description.

I don’t think it’s possible to make a conclusive case that geniuses have become rarer, a point Hoel himself concedes. The opposite may even be the case: There are so many highly capable scientists who would previously have been labeled geniuses that it’s harder for any of them to stand out. This is similar to the observation that while we still remember Andrew Carnegie for his enormous wealth more than a hundred years after he died, there are many billionaires today whose wealth exceeds his even after adjusting for inflation, but whose name will not be remembered by as many hundred years from now due to the number of rival billionaires they have to compete with for a place in magnate history.

Hoel observes that many of the geniuses of the 19th and 20th century shared one thing: They enjoyed one-to-one tutoring and some didn’t attend school as we know it today at all. By definition, tutoring doesn’t scale. At least there was no way to scale it until the advent of generative AI. Neal Stephenson’s science fiction novel The Diamond Age described an AI-powered tutoring device and how powerful it could be back in 1995, and creating something like it using generative AI has become the goal of many active in the educational space.

Andy Matuschak argues that this is misplaced:

I feel haunted by the Primer. I know it’s not what I want to build, but some part of my mind won’t let go of that vision until it has something else it can grab onto. In fact, I think my whole field is haunted by the Primer.

Matuschak’s argues the Primer isn’t a good model as it would be too prescriptive. It controls what the student learns, it poses only well-defined problems with well-defined solutions and it denies students’ responsibility for their learning.

While this rings true, I wonder how much responsibility for their lesson plans the geniuses of past centuries had. My guess is that at least initially, their tutors were setting the agenda.

The problem with AI tutoring may be a different one altogether. The effectiveness of tutors may have less to do with how they teach than with the emotional bond they forge with the student. The most effective tutors are those who are admired by their student and whom they wish to emulate. The student’s admiration for their tutor is the initial fuel that powers their desire to learn and progress, and only later, if everything goes well, will the tutor provide an off-ramp that allows the student to pursue their studies independently. This could be called the Narcissus and Goldmund model of tutoring, based on the novel by Hermann Hesse. It’s not something that AIs could easily achieve, given that the emulation of an AI may just not be something that anyone would want to aspire to.