This AI Stuff Has Gotten Good
This evening, with the help of Claude, I managed to:
- Port over all the old content from my broken blog at Reasonably.Social despite the source code being quite different between this platform and that one.
- Update the design and display of this site.
- Make the front page the latest post.
- Paginate and reformat the Archive page.
- Automatically generate tags for all the migrated content (the old site didn’t support tagging).
That took me about two hours after work. There are a couple of loose ends, like a handful of images and videos that didn’t port over correctly, and I haven’t worked on the mobile display at all. Maybe I’ll do that tomorrow on a lunch break. It won’t take long.
I’m not sure what we’ll do for a living a few years from now. Probably cruise vast deserts in muscle cars, battling over the last drops of gasoline.
You might be consuming YouTube videos and Bsky posts featuring so-called experts arguing AI is a hoax and doesn’t really work. A mere illusion. Nothing but a beefed-up ELIZA. Neurodivergent autocomplete at best. This is incorrect. AI denialism is an influencer grift. These conmen put you in harm’s way by refusing to acknowledge the position we’re in. A cyberpunk twist on antivaxxers.
I’m not making a case here that AI is a net benefit to humanity. Or that it’s sentient. Or that agentic software factories are a good idea. Or even that the circular economics behind the infrastructure aren’t untenable. I’m merely pointing out that AI is highly effective at most things you’d reasonably ask AI to do—and that list is growing quick. People who say otherwise are not so much mistaken as lying.
I work with it for hours on end, daily. Current AI implementations combining flagship frontier models and the latest harnesses far surpass anything HAL or Captain Kirk’s ship computer ever did on screen. We’re not quite looking at Data from TNG or Red Dwarf’s Holly, but things are headed that way along the continuum. And, frankly, we may actually be at Holly.
Perhaps the hardest thing to wrap one’s mind around is that much of what you learn about AI becomes outdated every six months. The capabilities are improving rapidly—and that’s just the publicly released stuff. Lorde only knows what’s going on inside the labs.