A week later, an automatic Google Scholar alert drops into my inbox. One of the professors I follow-Marchetti, from Sapienza-has a new preprint. "Hyperbolic Sheaf Convergence to Bridge Symbolic and Distributed Representations." I click the link and immediately forget whatever I was supposed to be doing. The abstract reads like it was written by a committee of mathematicians who hate each other, but the core idea smashes directly into the territory I've been obsessing over. It's literally my last three months of thought, but with real theorems and a bibliography thick enough to choke a horse. The panic is immediate. It takes exactly three seconds to spiral from "this is awesome" to "I have been scooped and should probably die." I pace the living room, consider rage-tweeting, decide against it, then try to find Marchetti's email. It's hidden behind a university paywall. I brute force it by guessing every permutation of his name and the Sapienza domain, then send a test email to each. Only one doesn't bounce. Dear Professor Marchetti,
My name is Alex Vukovic. I'm an undergraduate at UCSB, working independently on similar problems to those in your recent preprint. I have some ideas about how entropy gradients could drive symbolic abstraction in mixed architectures. Would you be willing to do a quick call or Zoom? I'd be grateful for any advice, or even a sanity check. Thank you for the amazing work.