Everyone is producing AI inbred slop

This newsletter is a collaboration with Kaguura Gichuru. He brought me the first draft, then we bounced it back and forth until we were both happy with it.

If you like it, make sure to subscribe to him on Substack.

Enjoy.

Between your ears is the most powerful supercomputer on the planet.

Eccentric Tech founders in flip flops are spending billions trying to build one that matches it. They need whole warehouses full of machines to power it. You need a $2 coffee and a slice of bacon. That’s it. Breakfast.

Yet, society is obsessing over how smart AI is getting.

Meanwhile, we’re handing the most powerful supercomputer on the planet over to a tool that can’t think, can’t wonder, can’t care. All to save ourselves forty-five seconds on a Tuesday afternoon.

We trained AI on the well of original human thought. Imperfect, biased, real. Now we’re training it on what AI made last week. Copies of copies, still imperfect, still biased, but without the humanity.

It’s not just slop anymore. It’s inbred slop. And that’s the DNA running in the background every time you use AI to write.

It’s like finding out your girlfriend’s grandparents are siblings. Then trying to start a family two days later. Gross.

People will tell you this is one of the big reasons to stop using AI.

To opt out of the most powerful tech in the world. It’s a valid solution. We won’t lie about that. But it’s not the one we’re betting on.

Because using AI well in the inbred slop era is a skill. A real one. With mechanics, levers, and practices like any other craft.

And almost no one is teaching it well. The loudest voices are either selling you the newest AI tool or screaming that AI is nuking everyone’s brains. Few are showing you how to use AI to write without losing your humanity in the process.

The small group of people who figure this out are going to look, in ten years, like they were playing real life Sims with cheat codes.

The first step is understanding the mechanics behind your own thoughts.

Evolution, by design, made you brilliant but lazy.

And if you’re going to fight thousands of years of evolution, you’re going to lose. Which is why so many struggle to use AI without losing themselves.

AI isn’t just another innovative technology. It’s the first tool in human history that lets you outsource your mind.

Not your consumption. People have been blindly believing books, propaganda, and authority figures for centuries. Your creation.

If you use AI, you might already be doing it. Because AI doesn’t create like you. It doesn’t think like you.

Actually, AI doesn’t think at all. It’s just really good at guessing. Literally.

When an LLM writes your content it’s running statistical probability at scale, analyzing patterns from its training data to guess the most likely next word in a sequence.

And it’s very good at this narrow skill set. Exceptionally good. But limited. It has never experienced what it’s like to be human.

It’s never been confused, or yearned to discover the meaning of life. You have. Because your brain thinks differently.

It compresses. Decades of pattern recognition turn into understanding, then knowing, and eventually instinct.

This is why learning to use AI responsibly feels hard and riding a bike feels effortless. Compression has turned thinking into being.

But compression, by design, is also what makes you lazy. The brain is metabolically expensive, and evolution shaped it to offload effort wherever it can.

Fire offloaded warmth. The wheel offloaded weight. The printing press offloaded memory. Every leap in human progress has been a story of handing some piece of work to a tool so we could free our minds. Create more room for creativity, critical thinking, and innovation.

Then AI came along. The first tool that completely offloads the mind itself.

Not a function of it. The whole thing. The wondering, the reasoning, the deciding, the connecting. And the brain, the lazy little optimizer that it is, is more than happy to accept.

Every time you reach for AI before reaching for your own thinking, you rust your cognitive rails. The you that would have emerged on the other side of that struggle fails to actualize.

What do you think happens when this isn’t just happening to you? When it’s happening to everyone in your community?

The professors at your universities. The doctors writing your prescriptions. The people running your government.

A fucking catastrophe.

“He who fights with monsters should look to it that he himself does not become a monster. And if you gaze long into an abyss, the abyss also gazes into you.”
— Friedrich Nietzsche

Most people who read this section will keep using AI exactly the way they did before. A small group won’t. The difference between them isn’t intelligence. It’s whether they let the next ten minutes change something.

The first AI models were trained on what we call the “Old World”. Millions of books, centuries of art, the messy and vibrant chaos of human discourse. Imperfect, biased, but real.

Now AI isn’t just copying humans. It’s copying itself.

On X. On LinkedIn. Increasingly on Substack. The slightly-above-average AI content is pulling hundreds of thousands of interactions.

Which signals to the next model that this is what quality looks like. Which trains more of it. Which gets more interactions. A feedback loop of averages. Mediocrity, on a loop, training the next round of mediocrity. You can guess what happens next.

Ideas, writing, and content become increasingly inbred, chinless, and fragile. A digital Habsburg Empire — a European royal family known for centuries of inbreeding.

On a cellular level, you become what you eat. On a relational level, you are (apparently) the sum of the 5 people you spend the most time with. But there’s a third part of this that’s missing.

Your mind. It takes the shape of what it consumes. And if you wouldn’t inbreed on a cellular level, what makes you comfortable doing it with your mind?

You’re not just watching this from the outside (honestly, neither are we). You’re inside the loop. Every time you publish AI-generated content into the feed, you’re feeding the next model.

Every time you let AI write the thing you used to write, you’re another data point in the closed system.

You’re not just consuming inbred slop. You’re producing it. Becoming it. Endorsing it.

And the worst part? The atrophy is invisible. Until the servers are down and you can’t think of something to write in your own mother’s birthday card.

This is a deep, deep hole we have dug ourselves into. And to get out of it, we’re going to have to do the work. The hard work.

Society will continue to atrophy. The slop loop won’t stop. But you can save yourself by becoming a sovereign individual who takes back control of your own mind. And uses AI to reclaim what was already yours.

This isn’t aspirational philosophy. This is neuroscience.

Share

Your brain can rewire itself. Even after extreme atrophy. What was lost can be found.

That’s the only good news in this story. But it’s enough.

Because right now, you’re copying with AI. And that’s not entirely your fault. We’ve already talked about the nature side. Evolution made you lazy. Your brain is metabolically expensive, and it will offload effort wherever it can. AI is the most efficient offload it’s ever been offered.

But there’s a nurture side too. And it’s been compounding on the nature side for decades before AI ever showed up.

A decrepit school system turned you into a cog within its greater system. It taught you the answer was more important than the question. It taught you to copy.

AI didn’t teach you to copy. The system did, decades ago. Evolution made you lazy, school made you a copier, and AI just walked in offering to do the copying for you. Of course you said yes.

You were prepared for this moment your whole life. By both nature and nurture.

So if you want out, you don’t fight AI. You don’t fight evolution. You go back to the thing that was working before any of this was done to you.

Alan Watts talked about how a child will dance to music. The parents see this and send the child to dance school. The child spends years learning how to dance “correctly.” Eventually, they become a professional dancer.

They dance again the way they did as a toddler. Only this time it isn’t called cute. It’s called a masterpiece.

That’s the move. Learn the rules. Master them. Return to play carrying everything you learned.

Your advantage against AI slop isn’t memory, processing power, or learning speed. AI has you (and us) beat on all three.

Your advantage is the thing you mastered at four years old. Before anyone told you what art was supposed to look like. Before you learned to copy one source and call it creativity.

A four-year-old doesn’t copy one thing. They take a dinosaur, a spaceship, a memory of their grandma’s kitchen, and a song they heard once, and they combine them into something nobody asked for and nobody else could have made.

That’s recombination. It’s our nature. It’s the part of you the system spent years training out, and the part AI is now finishing off.

The work is to get it back.

Not to become a polymath. Not to become some idealized creative genius. Just to reclaim the creative capacity you were born with and stop handing it over every time something easier is available.

That’s what the next section is for.

If you’re enjoying this newsletter, consider subscribing to this articles collaborator:

Kaguura Gichuru

When people first try to reclaim their creative capacity, they get it wrong. They start studying ten things at once. They take three courses. They pick up books from Nietzsche, Hawking, and Dr. Seuss. They tell themselves they have multiple interests. That they can’t just pick one thing.

This will be uncomfortable to hear. You likely do have multiple interests. But if you’re following this process, you’re not creative. You’re scattered.

But here’s what changes that:

You don’t pick multiple subjects. You pick one constraint big enough to require multiple subjects.

Da Vinci wasn’t randomly interested in anatomy and engineering and painting. He had one unified constraint: represent the world as accurately as possible. Anatomy was in service of painting humans. Engineering was in service of building structures that held. Optics was in service of how light fell.

The domains weren’t scattered. They were all pulling toward one thing at a higher altitude.

When we’re working on becoming better writers, the unified constraint is write better newsletters. But you can’t actually do that with writing alone. You have to understand how to think. Nervous system regulation, so you can stay connected to your intuition while you write. Marketing, so the work actually finds people. Graphic design and emotion, so the visuals carry the same weight as the words.

All of it pulls toward one constraint. One newsletter. One piece. One thing being made. Focused, but wide.

Same focus as the specialist. With the surface areas of a deep generalist.

Once you have that, the work splits into three moves.

1) Pick one real project. Make it the unified constraint.

Your brain needs a target. A specific problem to solve. That’s how it knows what to focus on.

But you must pick one big enough that recombination becomes necessary to finish it.

Writing a creative newsletter. Drawing your friend’s Golden Retriever. Producing a short, cringey, home movie.

This project is what challenges you to flex your creative muscles again. Without it, you’re either scattered or back to atrophy.

2) Produce first. Then use AI to tell you where the gaps are.

This is the move that determines whether you keep your creative capacity or lose it.

Before AI touches anything, solve at least 80% of it yourself. Write the draft. Curate the creative direction. Channel your inner Michelangelo and sit with the hard part long enough that something of yours actually shows up on the page.

Then AI tells you where the holes are. Where the logic breaks. Where your writing voice shrinks.

AI should be your editor. Not your ghostwriter.

Every time you let AI produce first and edit yourself second, you’ve handed back the part of the work that was building you. The minutes you save are the reps your brain was about to do.

3) Give AI the mechanical work. Refuse to let it touch the creative work.

There are two categories of work in everything you make:

The creative and the technical.

The thinking, the deciding, the connecting, the noticing. That’s the architect. That stays with you. It’s yours.

The formatting, the cleaning up, the converting, the repeating. That’s the engineer. That gets handed off. And thank god for that.

Get that line wrong and you’re back in the slop loop, just with extra steps.

Stay with this long enough and you’ll feel it. You’ll catch yourself noticing connections in your own work. You’ll start craving harder problems. You’ll start producing things AI couldn’t have made even if you’d asked it to.

Because the thing that makes them yours is the thing AI doesn’t have. The unified constraint. The lived material. The specific library nobody else has built.

The slop loop will keep running. Most people will stay inside it. They’ll keep chasing better prompts and newer models, getting duller without noticing, becoming the inbred content they consume.

You won’t.

You’ll be one of the small group who figured out what the tool was actually for. Who used it to do more of what should be copied, so you had more time to become the thing that couldn’t be.

Let AI do the dishes.

You paint the art.