My first question about language was visual. My thesis was on how text and image work together in Scandinavian silent film, two systems that don't simply repeat each other, but produce meaning in the space between them.
What fascinated me then was how the brain handles this: image and language are processed differently, and meaning emerges from how they meet. That question turned out to be less academic than I thought.
Because now we have AI that does something similar. Not neurologically, but functionally. Multimodal models are trained on billions of image-text pairs. They don't see or read the way we do, but they've learned what goes together, and they produce something that looks a lot like meaning.
Which raises the question I keep coming back to: if we arrive at the same result through completely different processes, what does that say about meaning itself? Is it in the process, or in the outcome?
Most people asking questions about AI are asking about capability. I'm more interested in what it reveals about us.
If that sounds like a useful lens, you're in the right place.