luke cassiano

I study cognitive science and neuroscience at Berkeley, and I build things — surf forecasting, multi-agent inference systems, AI tooling, the occasional overbuilt website. Most people file the science and the building into separate drawers. I never have; for me they’re the same drawer.

The thing I keep circling — in papers, at 2am, probably more than is healthy — is the seam where physics, neuroscience, and philosophy of mind run into each other. Not as rival answers to the big questions (free will, uncertainty, consciousness), but as different angles on the same one. I’ve spent a while convinced that consciousness is basically an estimator — a model the brain runs to guess at what’s actually out there, under a fog of noisy signal. Once you look at perception that way, it stops being a philosophy problem and starts being an engineering one.

There’s a funny tension in that: I more or less build inference systems for a living, and I still don’t think the ones we have are “thinking.” Mimicry isn’t qualia — a model regurgitating a dataset isn’t having an experience. They’re estimators, not minds. The distinction matters to me more than it probably should.

Which is conveniently the engineering I like. Surf conditions, subcontractor spreadsheets, an agent’s latent belief state — it’s all the same move: take messy, ambiguous, half-legible signal and turn it into something clear enough to actually act on. I don’t buy that rigor and feel are opposites, either. Reading a forecast, comping over a chord change, calling a hidden state — honestly, same instinct, different instrument.

Speaking of instruments: I’ve played bass for going on a decade (jazz, mostly), I surf, I snowboard, I make music, and I once spent ten weeks teaching a course on free will — which is about as on-brand as it gets. Improvisation, it turns out, is just inference you can hear.

That’s the short version. If any of it resonates — or you just want to argue about determinism — I’m around.