You've spent the better part of a decade pushing back against the idea that consciousness is, at root, a computational process. Has the AI boom made that argument easier or harder to make?
Both, in different rooms. With the public it has become much harder, because the implicit story — large model, looks intelligent, must therefore be on a path to consciousness — has the simplicity of a good lie. With my colleagues it has become much easier, because the moment you have to write down what you actually mean by consciousness, the computational story starts to creak.
You've said you're increasingly suspicious of the word "emergent." Why?
Because it is the word people reach for when they want the conclusion of an argument without the argument. It does real work in physics, where the bridge between levels is mathematical and explicit. In AI discourse it has become a kind of incantation. I would like us to recover the original meaning, or to admit we have lost it.