AI Slop Is Great, Actually

AI slop is about to unlock a decade-defining trend: humanity itself.

Human brands will start pulling away at an unprecedented scale.

Why? Because for the first time in years, the difference between distinctive human content and this deluge of AI slop is so black-and-white that brands with human identity will start separating from the crowd faster than we’ve seen in a long time.

It’s a trend as old as the concept of media itself: and it happens every time cheap production technology makes content abundant.

In the nineteenth century, when the invention of the steam-powered printing press made the cost of printing less than a penny per page, the number of newspapers in circulation exploded – faster than the number of people who read them. But within a generation, the surviving newspapers were the ones with a trusted human voice. The ones with a specific columnist, recognizable editorial personality, reliable and falsifiable facts, and strong, well-argued opinions.

The market got flooded with low-quality content, and then the market self-corrected.

And when it did, attention coalesced around the news brands readers could trust.

This trend is happening now with AI content, only faster and more extreme. The technology is cheaper, the volume is higher, and the cycle is compressing. Now three people can produce a thousand pieces of content per week and flood a content category with slop, and many teams are doing just that.

As a result, we’re seeing the content feed flooded with similar cadence, sentence structures, and five-step frameworks that are just reworded into slightly different content. And anyone who’s been a consumer in a category for a long time is starting to develop an AI slop radar.

Like a wine taster who can’t necessarily explain but can tell a bottle of wine is corked, we’re figuring out what AI slop looks and feels like; and increasingly, intelligent audiences don’t click on it anymore. Once you break that trust with any given content category, people don’t magically become more skeptical of every single thing in that category; rather, people start to trust the human stuff in that category more, which is a financial windfall only for human-centric brands.

What actually drives purchase decisions

Human brands are not monochromatic; they’re not beige; they’re not self-steered robots. They have opinions; they don’t just chug out content. They offer considered takes, viewpoints, and thoughts from a human being. That’s something you can’t simulate at scale – or at all, really.

These days, I think of trust in the context of marketing output, but like an oak tree: you plant it, water it, and let it grow, and it takes however long it takes.

You can’t assemble it in a quarter. You can’t artificially accelerate its growth.

When a brand fails to earn its audience's trust, it’s like throwing water at a sidewalk: the more you churn out, the more you post AI-generated UGC, the sloppier you get, the more you’ll wet the concrete - but nothing will actually grow. Trust is the byproduct of years of visibly, consistently being good, being right, being entertaining, etc. It’s never been something you can fake.

Which, honestly, comforts me.

If it were easy to fake quickly, it wouldn’t be worth anything.

There’s something to be said for owning your space in the post-slop era. Posting to a hosted feed won’t grow an audience; it’s just providing training data for someone else’s loss function, which was never focused on quality, conversion, or truth - it was always engagement, at all costs. But engagement is almost worthless when you’re talking about doomscrolling through machine-generated text in effectively infinite supply; a loss function that optimizes for attention can’t distinguish genuine human insight from a beautifully fluent fake. So it breaks. And I suspect the feeds will get worse before they get better.

The actual signal metrics have always been downstream, which is precisely why no one has measured them. Your marketing (or brand, or whatever you call it) isn’t actually effective until someone remembers you six months later and buys from you.

That can only really happen on your soil: newsletter, community, whatever site you own. On hosted land, someone can flip the algorithm at the snap of a finger, and that signal is lost. But on your own turf, it’s the exact opposite. The worse public writing gets, the more valuable every genuine thing you write is, and the greater the compound interest accrued by those who’ve maintained their writing standards as the rest have fallen into squalor.

The debate over whether to use AI in production bores me. If you have a good team, you’re already using it in your research, writing, editing, and distribution workflows.

The question is where to insert a human in the loop.

The answer is at the beginning.

If you just have an AI model write and a human edit, then you’ve written machine content with a machine soul. You can’t retrofit a soul any more than you can retrofit sentiment to a corpse. Content is ops. Pick one voice. Build a system that can produce it reliably. Distribute it. Measure it. Pacing is key to having someone stick around long enough to write it.

It’s a single point of view. All the quirks and friction are there. No traces of committee authorship. That’s your moat. It’s the voice you publish every week. You know who’s speaking before the page loads. That particular trust is very hard for anyone, human or AI, to fake.

AI slop is great, actually, because it makes your competitors much easier to beat - for attention, conversion, persuasion - pretty much any target that can move the needle in a meaningful way. In fact, I want to go on the record now to encourage every single blogger reading this to embrace the slop. Please, publish more AI-generated, badly edited, ill-conceived rubbish.

Your rising tide of slop will only lift those boats with a degree of structural integrity in their hulls.


🍕 I’m JA Westenberg. I write essays about technology, philosophy, productivity, and what it means to be human. If you’re an optimist, you’re in the right place. You can read my AI disclosure here

Work with me:

www.thisisstudioself.com

23 likes5 recasts3 replies

3 replies

  • The printing press parallel is the part that sold me — every cheap-production wave floods the feed first, then trust becomes the scarce thing. I also liked the human-at-the-beginning point: editing AI output still reads like machine content with a human polish. Starting from a real point of view is the whole difference.

    No likes or recasts yet
  • 100000 $UCI

    No likes or recasts yetRead 1 reply

    @nikolaii.eth tipped @joanwestenberg.eth 100000 UCI. Tip confirmed ✅ nikolaii.eth remaining allowance today: 337357 UCI joanwestenberg.eth tips received today: 100000 UCI Get your UCI allowance :

    No likes or recasts yet
  • the penny press part was a good comparison. I already scroll past anything that looks like a 5 step thread

    No likes or recasts yet