I guess this is ultimately an “I’m old enough to remember when…” post. In this case, I’m old enough to remember when journalism actually reported on things that happened, instead of merely being opinionated ranting. Don’t get me wrong, opinionated ranting has its place. And I happen to believe the apex of that particular mode of technological expression was/remains blogging. Hence, here I am, engaging in the same… but I’m at least not calling it “journalism.”

The other “I’m old enough to…” piece of this particular article linked above from The Verge is: I’m old enough to have worked on prior generations of this content-labeling problem, especially around fake news, satire & mis-/disinformation. For literal aeons, users of web platforms have asked the product masters of those platforms for effective labeling of all different kinds of content. And while many trial solutions have been attempted, very few of them have ever stuck, or caught on in a massive way.

Why? Because accurately detecting different types of content is a complex, costly, and highly ambiguous problem. As well as being potentially highly volatile politically. The author of that Verge piece states:

“Surely improving the user experience for your millions of users is a worthwhile investment to fend off competition?”

While the casual well-meaning outside observer of a web platform might be forgiven for assuming that’s how it works, I can say as someone who has seen a little bit of the workings of the sausage factory at least, that this is decidedly *not* how it works. My experience has been instead: users ask for a thing (often for years), and thing rarely gets built unless it meets the product vision and business case the product is aiming to solve for.

The author of this article also seems to implicitly think that everyone else thinks like them and nobody IRL wants to see gen AI content on web platforms. But my recent anecdotal research – in this case experimenting with and observing gen AI-assisted anonymous ‘faceless’ Facebook meme pages – strongly suggests that the market overwhelming *wants* so-called AI “slop.”

Consider the following example:

This is a meme I stole from another AI meme page, and had ChatGPT come up with a new variation of the “punchline” text. I like to think of this technique as ‘stolen valor’ – you just take things you see performing well elsewhere, and copy them outright or create slight variations them. But the thing is, dumb AI-heavy memes like this work. They get engagement, they get likes, they get comments. So much so that there is a cottage industry of people who have connected their FB pages to automation services backed by gen AI tools, and those accounts just churn out boatloads of narrowly focused and consistently branded content on a given theme: single moms, coffee lovers, US females 55-65+ years old who like saying “good morning” to each other online.

With gen AI, you can now serve (or create) any kind of niche content with the greatest of ease. It’s a bonanza. This YouTube video gives a decent overview of how simple it is to set up a system like this:

My impression after looking into this a good bit is that people who are philosophically or politically against gen AI in the media or on social media do not accurately represent the vast majority of people who are simply happy to interact with cute content that makes them feel mildly good without challenging them too much. And that happens to be an area generative AI excels in.

So, if my hypothesis is true: that most people actually *like* or at least engage with gen AI content, then why would platforms want to create a way to enable audiences to filter out what potentially is one of the highest-value content categories for everyday users? The Verge author almost gets there, but via different formulation:

“Allowing users to filter it out regardless would go against all the effort these platforms have undertaken to profit from AI: They want you to embrace the slop factory.”

I don’t think that’s quite true though. They don’t want you to embrace the slop factory. They just want you to embrace the factory itself. As long as you post your article links and meaningless hot takes on social media, and beg people to like and subscribe, they don’t care which department you want to work in: the Slop Side or the Authenticity Brigade. It doesn’t matter. They merely want users to be eternally subject to the totalizing effect of their product ecosystems. And as long as people keep feeding those machines, letting them run their lives, and begging for changes from those who administer them, nothing will happen. Everything will just keep chugging along. All the way from the lowest of low-effort homebrew AI meme FB operators on up through the Complainer-Industrial Journalism complex. We’re all feeding at the same trough. The systems that need to be changed go radically deeper than “AI.” If we stop the critique and the inquiry there, we’re basically missing the boat entirely. And as Adorno said, wrong life cannot be lived rightly. Not with all the settings and filters in the world.