Neuralink put out video this week of their first human patient — a guy named Noland Arbaugh, quadriplegic, moving a cursor and playing chess using a chip implanted in his motor cortex. I've watched the clip probably a dozen times. It's not what I actually hope for, longer-term, and I want to be precise about that instead of getting swept up. This is a real, meaningful medical device — restoring agency to someone who lost it, genuinely good, genuinely hard engineering. It is not "merging with AI" in the sense the company likes to gesture at in interviews. It's closer to a very sophisticated remote control than a bridge between minds. The actual gap between "read motor cortex signals well enough to move a cursor" and anything resembling what I privately hope this field eventually gets to is still enormous. I read the BrainGate papers a while back — that lineage of work goes back to 2006, a robotic arm controlled by implant, and it took eighteen years to get from that to this chess game. Slow, real, earned progress. Not a montage. Still. Watching it, I kept thinking about the whole-brain-emulation roadmap paper from Sandberg and Bostrom — read that a few years ago, actually, half convinced I'd never think about it again outside of an idle "wow, wild speculative field" way. Their own estimate was something like a coin flip on whole brain emulation existing by 2064. The paper's from 2008 — nine years old already by the time I first read it. I'll be in my sixties myself, if I'm lucky, by the year they guessed. That's a strange sentence to sit with. I don't know what I actually think PROMETHEUS's role in any of that eventually looks like. I think I'm just noticing, more and more, that the private hope I wrote about last summer isn't going away, and every real, incremental, unglamorous step the outside field takes — one chess game at a time — makes it feel slightly less like fantasy and slightly more like a long, unfinished argument I'm quietly still having with myself.