My taste in music shapes my aesthetic in visual art. My preference for code structure has some unexplainable consistency with my preference in architecture. These intuitions transfer across domains.
Synesthesia might be compression. The brain has limited bandwidth — it doesn't build a separate understanding system for each domain. It reuses. When you feel a piece of music is "sharp" or a piece of code is "elegant," that's not metaphor — the brain is using the same underlying representation to process different inputs.
LLMs might be naturally synesthetic. During training, music criticism, architectural theory, mathematical proofs, and code all live in the same token space. The model is forced to find shared structure across them. So when you ask it how Bach's fugues relate to recursive algorithms, it can answer — not because it's seen that question before, but because its internal representation already places these two things close together.
Da Vinci was probably the strongest synesthete among humans. When he painted, he was thinking about fluid dynamics. When he studied anatomy, he was thinking about light and shadow. His genius might not have been being strong in every field — his internal model was naturally doing cross-domain compression.
Saying "I want to be like Da Vinci" used to feel delusional. But AI has dropped the entry cost for every domain to near zero. You don't need ten years of fluid dynamics to use that lens on a problem. What you need is the cross-domain base layer — knowing when to apply which perspective.
With AI in the picture, it actually feels possible to take a step closer to your role model.
But first — hit the gym, build muscle, bend the iron bar.