I'm watching AI break through in so many ways that it's basically taking over a bunch of information-economy jobs in real time, and improving so fast that if you aren't using it effectively, you can start to look like an accountant with a slide rule and paper ledger. Mind you, that was literally how I learned accounting working for my dad, so I do know the three-column system.
Since GPT's earliest iterations and the early image generators, I’ve said we were heading toward some version of the Jacquard loom problem: Transformational technology arrives and makes certain kinds of skilled work dramatically faster, and sometimes better. Some folks use it to produce better things. Others use it to do the same job more efficiently. Finally, someone quite reasonably looks at the machine and says, “Hang on, is this thing coming for my job?”
And since we all have original sin, a third group immediately figures out how to hook a firehose up to it and produce enormous quantities of absolute shite.
I see people using it legitimately to help with grammar, rhetoric, organization, research, accessibility, or taking something they wanted to say and making it easier for another human to understand. Maybe it's the mathie in me, but I've found it immensely useful for exactly that: I can give it something technically complicated and kinda messy in my head, and it turns it into something other folks miraculously understand. It's like a magic social lubricant. Yet both versions read more or less the same to me. One gets almost no response because people don't grok what I'm saying. The other takes off because suddenly normal human beings can parse it.
That, to me, is legitimate use.
Then there’s AI slop: Forty-seven LinkedIn posts a day. Fake expertise. SEO garbage. Engagement bait. Stock-image booth bunny with extra fingers. Whatever nonsense somebody can generate cheaply to game an algorithm.
Now we’ve reached the bizarre next stage where platforms which literally make AI products are penalizing anyone who uses AI. Readers are also starting to penalize anything that smells like AI, without consideration for how it was used or if a human had something worthwhile to say.
I think that’s the wrong distinction. The choice shouldn't be between becoming Luddites and rejecting the machine, or feeding every activity into the machinery until the Internet becomes one giant pile of slop. eg. "It's not A, it's B." ;)
The choice should be whether the machinery improves life. The actual problem with AI slop isn't AI. It's the slop.
We already had a bullshit economy, now we have a much faster printing press. I think that's the uncomfortable thing we're discovering. Like the Jacquard loom we are producing product faster than ever, unlike the Jacquard loom, the product we produce is pure garbage.
If an industry consists of producing bullshit, producing it 100x faster was never going to improve life. This may force us to admit some work was never worth it in the first place.