I used to read self-help. Properly, for years. Books about discipline, then books about focus, then books about why the other books don't work. At some point I noticed the reading had become the activity, and the thing it was supposed to produce still hadn't turned up.
Keeping up with AI feels the same from the inside. New model, new tool, new technique, something worth understanding roughly every week. It's absorbing, it feels like work, it looks like work to anyone watching, and nobody has ever questioned an evening I spent on it. I notice how much I like that combination.
The mechanism is the same one. You get the feeling of the state without being in the state. Reading about discipline is the experience of being disciplined, at a discount. Studying the frontier is the experience of being ahead. Both come with a small hit of relief that quietly files the problem as handled.
They also share the lack of a finish line. Self-help is infinite by design and the frontier moves on purpose, so "caught up" isn't a state that exists and you can't fail at either. That's a comfortable arrangement if what you want is motion without exposure.
There's a supply problem sitting on top of it. Everyone gets the same model on the same day for the same price, so whatever it hands me for free it hands the next person for free. The frontier is the most crowded place on earth at the moment. I keep treating hours spent there as an investment when it's closer to standing in the longest queue in the building.
The complication is that some of it is the job. I build software and a fair amount of it is AI-shaped, so there's a version of this note that's just anti-intellectualism with a productivity excuse, and I don't want to have written that one.
The test I'd like to use: does something I'm building break if I don't learn this? If yes, it's research. If no, it's a documentary about my own industry. Self-help never had a named destination either, which is how it ran as long as it did.
I haven't been applying the test. I'm not sure I can from the inside, at eleven at night with a tab already open, which is the same position I was in with the books. I didn't solve it then. I stopped because something turned up with a deadline and a person attached to it, and that's the only mechanism that has reliably worked on me. It isn't insight. It's just having been needed somewhere at a specific time.
— Simon