September 12, 2026
I don’t think AI will destroy humanity in the next decade BUT I’m confident it’s going to make our lives much much worse. I know this because it’s already happening.
Today AI can effectively simulate almost every interaction we have with other humans. Every phone call, text message, every email, every blog post, every product recommendation, they can all be simulated at a level of fidelity so high you can’t be sure whether you’re interacting with a human or bot.
We’re building the world-wide equivalent of Google customer support.
Was your pizza order incorrect? Do you want to know if Macy’s has your size sneaker in stock? Did a cop arrest you because he mistook the 1 on your license plate for an I? Every one of these interactions can and will be handled by a bot, a hyper-logical entity that cannot (by definition) exhibit the flexibility and understanding of an actual human.
Computer said you need to go to jail, so you’re going to jail. No humans involved, nobody to talk to and explain that they read your plates wrong.
I don’t claim to have the perfect solution, but here’s two things we should do right now to stop the bleeding:
Most humans make less than 100 phone calls, send a few hundred emails a year, and send a few thousand text messages. Charging a nominal fee for all of this won’t affect anyone’s finances, even the most impoverished American. It WILL limit mass communication. I won’t hesitate to spend $0.05 to wish my aunt a happy birthday, but my stupid f*cking car dealership might think twice about sending me 5 emails a day.
As far as I can tell, using AI is a get-out-of-jail-free card for breaking the law. Hacking into a competitor’s servers to exfiltrate their data is illegal if a human does it, but appears to be consequence-free if you just let your AI do it.
We need to put a stop to this immediately. If an AI hacks into a server, or makes a race-based hiring decision, or commits defamation, a living breathing human must be held responsible. It’s self-evident to me that if we give humans the world’s most powerful law-breaking tool AND protection from consequences when they use it we’re providing strong incentives for some abhorrent behavior.
Yeah, we probably should regulate AI somehow? A lot of people are making a lot of noise about this, and they’re probably right. But, also, I have no idea how we’ll regulate AI.
« Desolation Wilderness | Home
9d52e36 · 2026-02-19 14:55 UTC