Hot take: The primary purpose and effect of "AI" isn't theft (though it is that) or making anything better/faster. It's about devaluing art and labor. The others are true, but they're a tiny detail to those people.

It's not about stealing art. It's not about making more art faster. It's about people who've never picked up a pencil or opened PhotoShop being able to look at a piece of art and go, "Pfft! Give me 5m and some DeepAI tokens and I can make that."

It's about people who've never invested time in learning to write well going, "Oh, books? Yeah, me and ChatGPT could write a book. That's nothing special."

It's about insecure white men trying to make up for that insecurity by building a helper robot instead of by actually learning to do the thing. It's about refusing to respect the people have invested in those skills.

Ya'know what people like Elon Musk hate more than anything else? Experts. Experts who actually know what they're talking about and tell them they're wrong. So they build a "superintelligence" which they claim knows all the answers because it's read the entire internet and then they ask it questions. But then, when it doesn't answer the way they want, they tweak the code and turn it into a yes-man, while still claiming that it's smarter than every expert out there. Now the "superintelligence" that's read the entire internet agrees with them so they can't possibly be wrong!

With software engineering it's no different. It's about CEOs and middle managers who haven't written a line of code in 10 years (if they ever did) looking at their engineers and going "Pfft! It's not that hard. You're just not working hard enough. Here! Have some Claude tokens. I expect you to double your output by next month." It's not their fault for refusing to spend enough to hire good people, making their engineers sit in traffic for 2 hours a day just so they can look out over their cube kingdom, having ridiculous expectations, micro-managing, or changing the requirements every other week. It's the engineers' fault for not using enough AI. They're still searching for the mythical man-month and some guy in a nice shirt sold them a brass lamp called "AI" and told them if they rub it just the right way a genie will pop out and make the man-month real.

It's also about mediocre engineers with big egos (if you've been around, you know who I'm talking about) being able to devalue the work of others. It's about them looking down at the hard-working junior and going, "I don't need you. I have a an AI. Me and my AI buddy will do better than me and you any day" so they don't have to invest time in mentoring the next generation of coders. And it's about them looking at the people who are better engineers (as good as they claim to be) and going, "Yeah, well... You'll see! Give me some AI tokens. Me and my AI buddy will code circles around you!"

It's about ego. It's always been about ego.

The good news, is that it will fail. AI is failing to deliver across the board and the bubble is going to pop one of these days. If you're investing time and effort into learning one of those trades, I don't think it's going to instantly become irrelevant. It may look different in 5-10 years but it'll probably still be there in some form.

The bad news is that the social and epistemological damage has been done. They've already accomplished their primary objective of destroying people's faith in experts and people who are masters at their craft. Even if actual artists or engineers know better, they've done a pretty damn good job in convincing your average Joe that those people are overpaid and over-respected. And that's going to take decades to rebuild.