Reflections
Personal reflections on presenting the In the Age of Intelligence lecture series at Saint Bartholomew's in January 2026.
What Worked
The arc of the entire series. Starting with the practical ("What is an LLM?"), moving to the philosophical ("What is Intelligence?"), and then landing on the provocative implications in the final lecture. Each built on the last. By the time we got to Dario Amodei's essays in lecture three, people had the foundation to engage with the harder questions.
What Surprised Me
I received positive feedback on Lecture 2, which I hadn't expected. I was worried it was too much to cover—deep sea vents, Conway's Game of Life, bacteria doing calculus, symbiogenesis, the whole sweep from physics to biology to AI. But people seemed to connect with it. Maybe the ambition of the argument was part of what made it work.
What I'd Do Differently
More structured demos with real depth. I showed some live interactions with ChatGPT, Claude, and Gemini, but I'd want to design demonstrations that really show off what these tools can do today—not just impressive, but practically useful in ways people could take home and try themselves.
What I Learned
Per usual, giving a lecture forces you to learn. The questions people asked pushed me further. Two insights crystallized for me: First, a better feel for how language translates into the internal models that LLMs use to predict. That compression from human meaning to mathematical space is fascinating. Second, how much the nuances of interaction—even just what's in the context window—can change the output dramatically, for better or worse.
Looking Forward
What a pleasure this was. It inspires me to look for other opportunities, whether in this space or far afield. Maybe I'll get a chance to teach a Vibe Coding class. Or finally give that set of lectures on Milton's Paradise Lost I've been wanting to do for years.