Listened to the whole episode and I highly recommend it. One reflection:

Carl Hendrick argues that AI and current technology can give us a way to analyze learning at a much more granular level. Right now, in many schools, whether learning happens feels chaotic and hard to predict. I know that’s often how I feel.

Carl’s argument is that either you believe that learning is amenable to the laws of biology, and we can improve our understanding of learning through better tools and better analysis, or you believe that learning is a kindof supernatural phenomenon, beyond our ability to comprehend.

That’s a fair comparison. I definitely see people in education who seem to believe learning is supernatural and I often find that unhelpful. I hope Carl and the rest of his team are able to generate some valuable insights for the rest of us.

I do think there’s a middle ground between the dichotomy Carl proposes. A better comparison might be meteorology. Weather is certainly amenable to physical laws, yet the complexity of those physical laws is such that long-term predictions are almost impossible. We can make decent predictions for tomorrow and three days from now. We can’t do much at all to predict whether there will be precipitation in 30 days. Weather is a physical phenomenon, but the complexity is such that we hit a wall with our ability to make solid, empirical predictions. I wonder if learning is the same: it is governed by physical laws, but the complexity is such that there are limits we won’t be able to surpass.

This doesn’t mean Carl’s project is futile. I would conjecture that our understanding of learning can go pretty far beyond where it is right now. I’m optimistic that there is much more to understand, and we can significantly improve student achievement beyond where it is right now. Still, I predict that some aspects of learning will remain mysterious and closed off to our best efforts.