Human slop

Two of my favorite authors are John and Hank Green, AKA the vlogbrothers. I’ve read most of the books they have collectively written. Twice each week I watch their channel and also Hank’s own channel where he talks about something on his mind and then plays Connections. I’ve even donated to a charity they started that built a maternity hospital in Sierra Leone. They are simply highly decent human beings that have graced the world with their presence.

But then Hank said a thing in a video that people thought sounded like ChatGPT. “Thanks for the pushback.” He was recording an edit explaining why his guest was right to question something he said. But some people thought, “Who is he thanking? That sounds like AI. ChatGPT likes to say ‘pushback’.”

It wasn’t AI. They were 100% Hank’s own words. But that didn’t stop people from publicly pooh poohing. No real research. No healthy skepticism. J’accuse! He’s the devil, now. They just repeat what other people have said as if it were fact. In other words, human slop.

Humans can generate slop, too. Especially when their incentives align with generating slop for an audience happy to consume it. Reporters will quote tweets like they’re news. Randos will record YouTube videos saying, “I’m disappointed in him – please like and subscribe if you’re disappointed, too.” Even the knitting community has jumped on the bandwagon because they didn’t like a knitting video he recorded a year ago. “Aha! That video must have been written by AI!” It wasn’t.

There’s human slop. There’s AI slop. I myself have produced both kinds of slop. And sometimes I even eat that slop.