← /contentslug: 2023-06-30-what-i-actually-want-this-for
date: 2023-06-30
title: what I actually want this for, if I'm honest
type: notebook entry
Dad had a scare at work this week — nothing serious in the end, some chest pain that turned out to be a pulled muscle, but the two hours between the ER call and the "he's fine" call were bad enough that I've been thinking about it since. He's fifty-three. Not old. Not young either, not anymore, not the way he was when I was a kid and fifty-three seemed like a made-up number.
I want to write down something I don't usually let myself say directly, even in my own notes.
Somewhere in the last year, the thing I actually want PROMETHEUS to eventually be capable of has quietly shifted. It started as "understand intelligence." Then "build a process that gets better at building intelligent systems." Both still true, still what I'm doing day to day. But underneath that, if I'm honest, there's a newer hope I haven't said out loud to anyone: that something smart enough could actually solve the parts of being human that have never been solved. Aging. Disease. The fact that everyone I love is on a clock I can't see and can't stop.
I'm aware of how that sounds. I've read enough to know I'm not the first person to want this — there's an entire real, serious field around it. Companies like Calico and Altos Labs throwing real money at cellular reprogramming. Sinclair's lab literally restoring vision in old mice by partially resetting cell age. The rapamycin data from the NIA's testing program is real and reproducible, not hype — a cheap generic drug adding meaningful lifespan in mice, and nobody's finished the human trial that would tell us if it does the same for people, because funding a trial for a drug nobody can patent turns out to be a hard sell. I know Kurzweil's predicted a lot of dates that came and went. I know Unity Biotech's senolytic drug failed its trial and the company basically folded that line of work. I'm not naive about the hit rate here. Most of it doesn't pan out.
But some of it does. And I keep thinking: what if the actual bottleneck on all of this isn't biology being unsolvable, it's that solving it requires more careful, parallel, well-evaluated experimentation than any human research group can currently run. Which is, categorically, the exact kind of bottleneck PROMETHEUS exists to work on.
I haven't told anyone this is part of why I keep going some nights when I'm exhausted and would rather sleep. Writing it here instead, where at least future-me will know I meant it honestly, whatever comes of it.