Matt

Well, they’ve finally done it. Through raw force of will and more money than god, the nerds have finally done something that will “make the world a better place.” By that, I mean they’ve pooped out something will finally change the world in big ways, for better and worse.

As an early AI critic, I say this now, because the spicy autocomplete app finally excreted some text I could use, in the form of computer code that actually compiles. The AI hype prophesy made only five short years ago, that “it’ll be better tomorrow,” has finally come true.

This isn’t a sudden change of heart; a moment when I get down and pray to the matrix-math machine so it might take pity on my mortal soul the day it starts flying killer drones. No, I just saw the stupid clanker mimic human language into usefulness, and ya know, it got me thinking our illustrious techbros are onto something.

So what are we really talking about? Particularly, it’s the thing that every marketer calls “AI,” when they really mean GenAI or generative AI, or more specifically Large Language Models (LLMs) and the garbage heap of products powered by them that can excrete text, images, audio, and so on for us all to feast upon (collectively, “slop”).

It’s an important distinction if we’re getting deep into it, but for simplicity here, I’ll follow our intrepid marketers and just call it all “AI,” too.

Vibe coding

So today, the AI helped me “vibe code” a few apps into existence, and to my surprise, it did well. Finally.

Back in 2023, I’d heard from all the prompt jockeys (first descendents of web3, next in a proud line of cryptobros) that I’d be “left behind” if I didn’t get on the AI train back then. Well it turned out the train is still in the station in 2026, and with like two google searches I learned about all I needed to make it work well.

So then I told the chatbot what to do and it pooped out a website that… looked pretty good! (Obviously, I’ve graduated from a lowly prompt jockey to a Prompt Engineer! 💪💪)

Then I went back and forth, telling the machine to fix its bugs, and it eventually predicted-out a functional web app written in my language of choice.

It compiled. It didn’t assault my human ability to see. Suddenly, I felt the raw power of the el-el-em surge through me — I was 10X codemaxxing, mogging my brogrammers of yore still toiling around, writing each line by hand like peasants.

So I kept going. Next, I told it to make a little command-line utility for me. I found some advice on reddit, which said to use “plan mode” first, so you can check the dumb computer’s work. So I did that.

Within a few minutes, it evacuated a fully-functional app onto my screen, all in one go. The ramifications of this hit me in the face like a bag of wet hot dogs as I was rendered blind by the future suddenly dawning on me — I am obsolete. I am nothing. From here on out, it’s only vibes and loving the machine. I love the machine!!1!

But seriously…

It’s nice that the dumb toaster works now. I mean, using it for these low-stakes projects was actually enjoyable. I got to be dumb, lazy, and brainless, and the computer slopped out a whole bunch of sloptastic work for me. What’s not to love?

But while it seems useful for coding things that don’t really matter, the experience hasn’t changed my distaste for the thing. In fact, as my slop-coded apps grew in complexity, my rosy view of the thing grew dimmer.

I quietly ran into some of its inherent limits, and simply had to guess when it would screw up next, all while I grew more (falsely) confident in its output. Most bugs didn’t hatch from bad code, which I always reviewed, but in more insidious ways that would crawl out when I did more extensive testing. It crapped out functional code that looked fine, but for example, every once in a while would include silly little logical errors that any competent human wouldn’t have written in the first place.

Basically, as much as I’d love to, I won’t be lobotomizing myself to forget two decades of programming experience just so I can rent out a brain from our new AI overlords.

Am I obsolete?

A common refrain is that “coding is dead” — a view earnestly declared by AI adherents every year, but one that apparently we’re finally getting to, at least in practice.

We still need code, and we need people who understand how it works. But at this point, I don’t think it matters whether the tech does what it's promised. Right now, there are enough executives in corporate America just itching to lay off 40% of their workforce and then make the rest use the dumb toaster, regardless of if it makes toast. (Then, of course, they’ll expect all employees to “10X their output.”)

It’s not just the private sector, either — there are enough people in government blindly buying the hype (or looking to profit from it) that they’re joining in on the fun, cramming AI into all the holes it doesn’t belong in. And people will just have to deal with it, as more machines that can't be held accountable affect their very real lives.

So in this way, it’s clear the sentiment “coding is dead” is more of a declaration of intent than an observation. We see this when AI’s most fervent bros are downright gleeful to declare the “death” of movie production or writing or making music, most often so they can be the ones to control how all this culture gets made, all within their little AI app, naturally.

These crucial human creations and pastimes aren’t exactly dead. But a lot of people in the world sure are in a rush to kill it — usually so they can have it for themselves.

It’ll be different this time, bro, I swear

No matter which direction it all goes, we are at a turning point — most interestingly (not really) of putting all our faith and future into a handful of tech companies again.

It’s nice that LLMs can stochastically generate some useful code and speed up my development process. But right now, there’s a whole lot of CapEx turning forests into data centers across the country. So when exactly do these AI companies make all that back? Then what does a poor, lowly sloplicker like me do when my little chatbot friend suddenly costs 10X as much?

It’s the age-old story you don’t even have to go to business school to understand: every tech startup heavily subsidizes their product in its early days, making it free or cheap, to try and capture the market before their competitors do. This can go on for many years, as the company lights a whole bunch of investor cash on fire. But one day the bill comes due, and as a user, suddenly your AI girlfriend is a lot more expensive, and ignores your calls until you upgrade to the Deluxe plan.

Perhaps most of all, we can’t forget that some of the biggest companies building our AI “future” right now are run by the very same people that brought us the ad-based surveillance economy we swim in today. As much as we all love targeted advertising and being harvested like cattle for our attention, do we want to also give them a direct view into our every conversation that the average user perceives as “private,” between just them and the magical robot?

The importance of shitting on AI

This brings me to my final point: the importance of mockery.

At this point, I do believe our AI-generated / -powered / -mediated future is “inevitable,” just as our formidable AI bros foretold. The hole-cramming will continue until we stop calling the corporations “Microslop”, apparently.

Ah, but “Microslop” is the point.

These companies gave us a great gift by rolling out their minimum-viable research projects and proceeding to force-feed us ✨ sparkles ✨. If they didn’t, we wouldn’t already have such rich language to describe this world-eating, anti-human tech that threatens to further enshittify our world. For example, “slop,” as a unique AI product, was Merriam-Webster’s word of the year in 2025. “Hallucinate,” that cute euphemism for “outputting bullshit,” was Cambridge Dictionary’s word of the year in 2023. This gives me hope.

And I believe the mockery should continue. Because while we’re all dazzled ✨ by how the clanker can clank code, poop out emails, and excrete surreal audiovisual material, it’s also misdiagnosing people, substituting humans in an already lonely world, and creating deepfakes. It’s being made mandatory in the workplace, despite its many fundamental shortcomings as a technology. (If it was so revolutionary and beneficial to people, wouldn’t people just use it on their own?)

But there is a wonderful cultural tradition on the internet, a final bastion born from decades of tech “innovation,” where we get to collectively laugh at an invention that is idiotic on its face, from NFTs to the Metaverse. Many people have rightly had a similar allergic reaction to AI, and perhaps even more so because of its sheer arrogance — the way it was thrust upon the world, built and extracted entirely from the stolen work of humans, and with literally only one value proposition: eliminating the very jobs people need to live.

In the face of so much power and capital behind this project, mockery is an important check. And so, I say anyone with any sense left should employ exactly this, for as long as we can, as this technology threatens to rule over more of our lives for the foreseeable future. Against an unstoppable effluent of hype and slop, it may be all we have.