AI is the solution to all of life's problems

I listened to a recent Hard Fork podcast about Google's AI debacle with Gemini, and decided to give ChatGPT a(nother) try for a fun bit of self-torment. 

Like a lot of people, I played with AI a bit (MidJourney, Bing, DallE, and ChatGPT) but stopped once the novelty wore off. AI is a magical solution in search of a problem. I hear it's useful for programming. How about for writing?

To find out, I "asked" ChatGPT to summarize my novel City of Orange. (Putting "asked" in quotes because I hate anthropomorphizing AI. Gives it undeserved credibility. Then again, we humans anthropomorphize goddamn every little thing around us, probably because they love it so much.)

um, no

ChatGPT is so wrong it's funny. I've never written about a Jin Park, Orange, CA, or eSports. The other parts (immigrant experience, self-discovery) are pretty good...for my first, *entirely different* novel, Frankly In Love. I guess my next step would be to fine-tune my prompt til I get the result I want? How long will that take? Longer than writing a summary myself? (Yes.)

(I also asked ChatGPT to convert work in progress from first person present to third past, and it did, but it also changed the story itself. What an amazing timesaver!)

right, no, left, aaaa

I complained to my very smart brother about how ChatGPT hinders more than helps my line of work. He's experimenting with AI to streamline his content management workflow, which is a much better use case. 

"You're still on version 3.5?" he said. "Sign up for 4.0, it's way better."

He means well, but that sounds like, "The wheels on the 2023s sometimes turn left when you mean right. Upgrade to the 2024 model." Which begs the question: why is the crappy 3.5 version still even available when it so badly hallucinates? Uh, I mean, outputs such wildly inaccurate probabilistic outcomes? (I should just give up and anthropomorphize.) 

your onus, not mine

ChatGPT is aware of its limitations and follows the age-old Internet custom of placing the onus of safety entirely on the user. Don't share sensitive info (we'll just copy it, and who knows where it'll wind up?), check your facts (because we never will!).

is this fun?

Here's the thing about ChatGPT. A) No one asked for it—it's Silicon Valley bros sticking tech where it don't belong. 2) It won't improve productivity, because everyone will have it. And if everyone has a leg up, don't they all just wind up the same height anyway? III) It's creating a market for "query composers", which sounds more like a character in Westworld than a real job. And finally D) it is quickly building a body of spectacularly mediocre "content" (that word) which it will then train itself on, a dog eating its own excrement, spreading a sewage lagoon of words that will degrade the aesthetic standards of the humans consuming them.

I'm being dramatic. Mostly I think AI is a big huge silly distraction from real problems. A shiny new toy to play with and bet VC money on. It's fun! 

Way more fun than finding cancer cures, ending racism, providing housing, education, healthcare, universal basic income, etc. I wonder: is it possible to get VCs as excited about those problems as they are about giving schoolkids convenient ways to get stupider by cheating on their writing assignments?

calm down, dave

Why does AI make me so crazy? Because I'm a creative person. And guess what? You're a creative person, too. You are all creative, whether you think you are or not. (Subject for another time.) Anyway, it bugs me that AI is touted as some kind of ultimate artistic superweapon. It's not. At best, it's a production tool on steroids (eg. content-aware fill in Photoshop) whose promised gains are mostly imagined. AI cannot think, therefore it cannot be creative, therefore it is boring, therefore it is not any fun.

For me, fun is learning how to write, draw, and play music. Fun is doing this stuff with people in real life, too. People will always be more fun than technology. And interesting people are simply everywhere! So why bother with shitty robots?

about that letterloop thing

Quite a few people said yes to Letterloop, so let's give it a whirl and have some fun! I'll think of some goofy questions to ask, and then we'll go from there. In the meantime don't be surprised if you find an unfamiliar email from Letterloop pop up in your inbox, okay?

Thanks for reading!

— dave

David Yoon is the New York Times bestselling author of City of Orange, Version Zero, and for young adults, Super Fake Love Song and Frankly In Love.

He’s also the co-founder (with wife Nicola Yoon) of the Random House imprint Joy Revolution, which publishes love stories starring people of color.