the corner office

The FT has been running a series of articles investigating AI and its effect on work, education and society. Apart from nitpicking over their blanket use of the term AI, they've been interesting to read, partly to understand how the world is being affected by these technologies, but also to understand, and occasionally to despair, about how different people think about these technologies.

A recent article, Chatbots in the classroom: how AI is reshaping higher education (paywalled) discusses the reality of LLMs in modern education: some benefits, a lot of blind cargo cultism, and the inevitability of students relying more and more on LLMs to do their work and thinking for them. And this while early studies start suggesting that LLMs erode the ability of their users to think critically.

It had one quote from an AI proponent which stood out for me me:

Social media marked the end of the era of boredom; AI is the end of the era of confusion.

Both parts stood out. First, boredom. A line I roll out regularly with my sons, met often with eye-rolling and occasionally with gnashing of teeth: boredom is one of the greatest gifts you can give to a child. I don't mean locking them in sensory deprivation chambers or anything, but just that an excess of screens and a constant need for engagement isn't healthy. Boredom is when creativity sets in, when the mind is set free to think and dream new things. When you look around and see possibilities in everything, when an old toy comes off the shelf or a cloud looks like a castle, or a jumbo jet, or you reach for a pencil and paper. Should it be any different for students, or for adults?

The second, confusion. I immediately thought of an old clip of Richard Feynman, which I first saw years ago:

When you're thinking about something you don't understand, you have a terrible uncomfortable feeling, called confusion. It's a very difficult and unhappy business. And so most of the time you're rather unhappy, actually, with this confusion, you can't penetrate this thing. Now, is the confusion because we're all some kind of apes that are kind of stupid, ... trying to figure out to put two sticks together to reach a banana and we can't quite make it, the idea. I've got this feeling all the time that I'm an ape trying to put the two sticks together. So I always feel stupid. Once in a while though, the sticks go together on me and I reach the banana."

More seriously, I think confusion is good. Mistakes are good. Confusion and struggle are half the journey of learning and understanding. I'm not always as good at applying it as I should be, but probably the best and most common advice in the prefaces to my various textbooks is this: do the exercises yourself. Struggle. You learn nothing by looking up the answers.

I'm not a Luddite when it comes to LLMs, but there is always opportunity cost. Much like the students in the article: those who use LLMs to point them in the right direction, and then undertake the journey themselves, are the ones who benefit the most. The ones who use LLMs to avoid going anywhere at all, gain nothing.

And anyway: rather for me the humility and uncertainty of a confused Richard Feynman, than the confident prognostications and certainty of AI visionaries.