Is This Anything? 7
People talk about how AI will enable "the first one-person billion-dollar business"; I find this very dumb and it aggravates me. (I can't fully explain why I find this annoying, out of all the dumb things in existence, but so it goes).
Basically, "one person billion-dollar business" feels like a fact about your legal setup and contracting decisions rather than a real fact about your business.
I have no idea what her business situation is like but I'm pretty sure JK Rowling could have been a one-person billion-dollar business if she outsourced everything to contractors. Why should I care if she hires a full-time employee accountant vs signing a non-employee contract with an accounting firm? I'm sure the "one-person business" people are pointing at something, but I don't know what it is.
There's an economic model I've had vaguely in my head for a while, but which I've never really concretized, that's roughly like this:
Once upon a time, if you lived in a village and played the fiddle, you had a local captive audience that couldn't hear music easily except from your local band. You could get a ton of Live Music Performance Reps in with an appreciative audience, and after a while you'd be good enough to go to the nearest town and get more reps with a bigger audience, and eventually you could move to the city and become a fiddle superstar. I'm sure you had lots of other problems but in terms of your fiddle–playing this seems really ideal to me.
In the modern day, everything is different. People can listen to world-class fiddlers on their phones all day, or watch videos of fiddle concerts by the world's best fiddlers, and with improvements in transportation (and incomes) even hearing those musicians live is infinitely easier than it was a hundred years ago. And for consumers this is in some sense Good, but it also creates a problem of "how will our next generation of champion fiddlers learn to fiddle?"
Anyway. It strikes me that plausibly AI will create this problem in a whole bunch of new domains. I know there's a future where this is irrelevant because even the best fiddle-player is supplanted by AI two years from now, but suppose we're in a world instead where AI is better than 95% of people at any given task but the best human practicitioners are still better. How do those best human practitioners ever get to be best in a world where their many years of learning-curve are not appealing to any audience because AI provides a better version? Yes you can practice alone in your room for years until you show the world anything, but a) this is very unrewarding, and b) there's some things you can only learn by doing them for other people. I worry about this.
One possible outcome of AI (again, in the short term) is that it's a massive boon to "solo creators" who aren't that great at getting along with other people. In the old days, if you wanted to learn songwriting, you had to be good at songwriting but also either 1) good-enough at playing music to perform your songs, 2) good-enough at friend-making to find someone who would sing your songs. To the extent that AI creates a good-enough complement to all kinds of skills like this, the benefits accrue disproportionately to people who aren't good at finding human partners to work.