Look at them go. Two robots running a 100-m sprint in Beijing. Side by side, legs moving in ways that would have looked like science fiction just 5 years ago, one of them pulls ahead, crosses the line, and wins. In 2026, a humanoid robot can outrun every human being who has ever lived. That is real. That happened this week. And if robotics is moving in this direction now, I generally look forward to seeing what kind of robots exist in 2030.

What is happening right now is that artificial intelligence can do things like this, things that make you stop and think about just how far the technology has come. But in the meantime, most of the conversation about AI has been hijacked by a single word, slop. Today, I want to talk about what that word actually means, who benefits from you using it, and whether the panic about AI-generated content is solving the problem or just creating a new one.

In June 2026, something happened that should have been front-page news everywhere. For the first time in the history of the internet, human beings became a minority on the network they built. Cloudflare, which manages traffic for millions of websites, reported that bots now generate 57.5% of all webpage requests. Their CEO had predicted this crossover point wouldn't arrive until the end of 2027, and it came more than a year early. The dead internet theory, which is the old conspiracy that most online activity isn't actually human, is no longer a conspiracy. It's a measurement dispute. Cloudflare says the crossover happened in June. Thales, which is a French cybersecurity firm, argues it already happened back in 2023. They disagree about the timing, but neither disagrees that it happened. And the new bots are just scraping content. Traffic from AI agents that take action on the web, I'm talking clicking links, filling out forms, completing tasks, grew 7,851% year-over-year. Seriously, 7,851%. That's insane.

On YouTube, the platform is visible enough that the platform has started demolishing its own ecosystem. In January 2026, YouTube permanently terminated 16 channels with a combined 35 million subscribers and 4.7 billion lifetime views. Some operators had been launching 115 new channels in a single day, scraping viral videos, spinning scripts through AI, generating visuals, and uploading six to seven videos per channel per day. It sounds a bit like an automated digital sewer line running at full pressure. A Kapwing study found 278 channels producing nothing but AI-generated content had collectively amassed 63 billion views and an estimated 117 million dollars in annual revenue. On YouTube Shorts, roughly one in every two recommended videos on a fresh account is either AI-generated or what the platform itself classifies as brain rot. When a multi-billion dollar tech platform's official internal categorization for its own algorithm's is literally brain rot, you know we're no longer steering the ship, we're just trapped in the engine room altogether. And on LinkedIn, our favorite of all, lol, over a million people have clicked a seems like AI slop button in less than a single single month. An AI detection company found that 41% of LinkedIn long-form posts were flagged as fully AI-generated. The flood is real and the people are drowning in it.

But here's what I find interesting here. Everyone calls this stuff slop. A million people click that button. Merriam-Webster named it their word of the year. Countries are banning AI content from their charts. And yet almost nobody has stopped to just ask a fairly important question, what does slop actually mean? Merriam-Webster's official definition as of 2025 is digital content of low quality that is produced, usually in quantity, by means of artificial intelligence. Word of the year, very official stuff. The term itself started as slang for soft mud in the 1700s, became food waste in the 1800s, and now describes whatever AI puts on your screen that you didn't ask for. Essentially, we've come a full circle. Now, from feeding leftover scraps to pigs in 1840 to feeding synthetic algorithm slurry to humans on TikTok in 2026.

But let's glance at what that means really. We're nothing if not nuanced on this channel. Low quality and by means of artificial intelligence, those two things have been welded together as though they're the same condition, as though the AI part is what makes it low quality, as though the tool used to produce something determines its value. Generative AI did not invent junk content. Content farms existed long before ChatGPT. Clickbait existed long before ChatGPT. Fake reviews, scraped pages, template videos made for advertising sites, all of this existed before a single large language model wrote its very first sentence. What AI changed was the economics. It made producing garbage virtually free.

This is a distinction that matters a huge deal because when you say the problem is AI, the solution looks like banning AI or labeling everything AI touches or building detection systems. But when you say the problem is cheap mass production of garbage incentivized by ad revenue and engagement algorithms, the solution looks completely different. It looks like fixing the incentive structures that made the garbage profitable in the first place.

Think of generative AI as an amplifier. An amplifier doesn't have a sound of its own. It takes whatever signal you feed it and makes it a lot louder. Feed it a beautiful melody and you get a concert. Feed it a static and you get a very loud headache. If you give an experienced senior developer access to an AI coding assistant with unlimited tokens, they produce better code because they have the judgment to direct the tool, to catch its mistakes, to know what good architecture looks like before the first line is written. But give the same tool to somebody with less experience and they produce more code faster, but sloppier and harder to maintain. The tool didn't change, the human input did. And this applies to everything: writing, research, advertising, music. The AI amplifies what you bring to it. If what you bring to it is a bot farm designed to launch 150 YouTube channels per day, you get slop. If what you bring is a decade of expertise and a genuine argument, you get something entirely different.

There are two sides to every coin, except I suppose in quantum mechanics where the coin can be both sides simultaneously until you observe it, at which point it becomes whichever side annoys you more. But the point stands anyway. Nuclear energy, for example, can power a city or contaminate one. Code can hack or it can provide functionality. The capacity for misuse has never in the entire history of human technology been a sufficient argument for abandoning a tool. It's an argument for building safeguards around it. We didn't ban electricity just because somebody used it to make a singing animatronic bass mounted on a living room wall. We just learned to live with the regret.

Now, I have to be clear about something. I am not defending the bot farms in any way. I am a YouTube creator myself. I understand the game that they're playing. I understand the economics. And my advice to anyone running that operation would be exactly the same. Stop. Not because it isn't profitable. Clearly, $170 million a year suggests that it is. But because it's building something corrosive. It is eroding the very ecosystem that makes the profit possible. A bit like a termite eating its own house.

YouTube's own head of trust and safety, Matt Halprin, tried to draw the line. He said, "AI can actually allow people to make a lot of videos. Sometimes those videos are great and it really enhances creativity and you can create a higher volume of high-quality content that we want to encourage and have in YPP. But that exact same new tool can allow you to make lots of videos really quickly that are very similar. They're very generic and don't really have a narrative arc and don't really show your creativity. So, the same technology really enables great stuff, but it also enables stuff that's kind of content farming." And that distinction, in my opinion, is exactly right. But the difficulty of enforcing it tells us something important about why slop as a blanket term is so appealing. It's much easier to condemn a category than to evaluate individual cases.

But the fact that some people misuse AI to produce industrial quantities of garbage should not, cannot, mean that everyone who touches AI is producing garbage. An expert in any field using AI to refine an argument, synthesize research, or make a representative thumbnail is not the same activity as a bot farm uploading seven videos a day to 150 channels. The Merriam-Webster definition doesn't distinguish between these two things, and that is a problem because right now the word slop is being used as a weapon against both.

Before I show you exactly how much economic damage the slop panic is doing to everyone, including people who have never touched AI in their lives, let me tell you about today's video sponsor. This video is sponsored by Incogni, a service designed to help you take back control of the personal information that's floating around about you online. If you've ever Googled yourself, you may have been surprised by just how much personal information turns up. And that matters even more now because AI has made scams frighteningly easy to personalize. Your name, address, phone number, workplace, relatives, and other personal details can all become raw material for impersonation, fraud, and increasingly convincing AI-powered scams. Incogni helps remove that information from data brokers and other sites that expose it. Their automated removal engine handles requests and follow-ups across 420-plus sites, so you don't have to spend hundreds of hours trying to do it yourself. And with Incogni unlimited, their exposure scanner goes looking for additional places your personal information appears online. If something eligible turns up, custom removals lets you send it directly to Incogni's privacy specialist who handle the removal process for you. Because when scammers have less information about you, they have less to work with. Go to incogni.com/elai and use code elai to get 60% off an annual plan.

Right, back to the damage, which is profoundly economic and measurable. On CGTrader, which is a long-established marketplace for 3D assets used by game developers, film editors, and 3D printing enthusiasts. AI-generated models now represent one in every six uploads, but here's the interesting part. They account for just $1 out of every 90 in revenue. Buyers overwhelmingly refuse to pay for them. The company's own report basically said buyers are voting with their wallets and AI-generated content is struggling to compete. So, buyers can tell. They can smell it from a mile.

But, what happens when institutions try to draw the line for them? In Australia, an AI-generated variation of Madonna's Like a Prayer spent 16 weeks in the top 20, peaking at number two on the Australian singles chart. The response was a blanket ban. The Australian Recording Industry Association announced that wholly AI-generated tracks are no longer eligible for the ARIA charts or awards effective this Friday. Music must be substantially human-made to qualify. And sure, the charts are one thing, but the backlash didn't stay in music. I'm telling you it's everywhere.

In advertising, it has been swift and very expensive. Coca-Cola released an AI-generated holiday campaign, got hammered for it, and then did the exact same thing the following year with animated polar bears. Same trucks, same backlash, louder. There is a special kind of corporate hubris, I think, in watching an entire internet tell you this looks creepy, and your executive team responding, "Right, sure. But, what if we make the uncanny polar bears do it again next year, right?" Alex Hirsch, the creator of Gravity Falls, suggested Coca-Cola's iconic red came from the blood of out-of-work artists. McDonald's pulled an AI-generated ad after viewer outcry, and the interesting thing is that audiences didn't need a media critic to explain what was wrong. They just knew. The vocabulary appeared almost spontaneously. Soulless, uncanny, digital slop, as though the public had been waiting for a way to describe a feeling they recognize but couldn't yet name.

But then the pendulum swings the other way entirely. Dove pledged never to replace real people with AI in advertising. Brands like Aerie, Equinox, and Almond Breeze ran explicitly anti-AI campaigns. Aerie's no-AI positioning tracked to a 23% sales lift. Made by humans has become a selling point printed on the label like organic or fair trade. And this is where it stops being an AI problem and becomes everyone's problem. A 2025 survey by Raptive found that consumer trust drops approximately 50% when content is merely perceived as AI-generated, regardless of whether it actually is. Perception alone drives the collapse. It doesn't matter if a human wrote every word. If the audience suspects AI was involved, the trust evaporates.

I am regularly asked whether I'm AI because there's AI in the name of the channel despite having made videos for several years now, and despite having made a video discussing this specific issue, showing a video clone of myself and the real me to show the difference side by side. AI is in the name because we discuss AI. The background remains simple because I need to travel occasionally, and I'd like to show you the sources to my claims on screen without them hiding some painting or random thing in the background. I do have hands. You can see them here. If you still think that I'm AI, please go check out the Bohemian Rhapsody AI sung in a by me in six voices. If you can generate that with AI, I don't know. Congratulations. You seem to be better than all of us.

A separate study, however, found that 59% of voters believe that content they encounter online is often shaped by AI. And whether they're right about any given piece of content is almost besides the point, the belief itself is changing how people decide what to trust. It's a bit like the monster under the bed, right? You don't need to check if it's there, you just pull your feet up. A real photo gets dismissed as AI-generated, human-written copy feels automated, an authentic voice gets lost in a feed full of synthetic ones. The slop has poisoned the well, the human well, for everyone, including people who have never opened ChatGPT in their entire lives.

And I don't entirely blame the public for the backlash, by the way, even though I'm regularly being called AI, and often those people intend to insult me. I mean, look at what the AI industry has done with its own messaging. When these companies were trying to raise billions in funding, their pitch to investors was that AI would replace workers at scale. That was the selling point. Half of all entry-level jobs gone, massive efficiency gains, do more with fewer humans. They said this loudly and very repeatedly, I'm sure you all remember, because it made the stock price go up. And then they acted surprised when the public turned hostile. I mean, if you spend two years telling people that your technology is going to take their jobs, you cannot then be baffled when they don't want to engage with that technology. The economic anxiety and precarity are a living reality for most humans on this planet, and the companies that created them are now wondering why the vibes are off. The vibes are off because you told an entire generation that their labor was about to become obsolete, and you said it like it was a feature. What exactly did you expect other than VC funding?

Anyway, where does that leave things? There has to be some solution to this, a way forward for all of us here.

First, I don't think the Aria ban is wrong exactly, but I think the framing might be incomplete. AI generated music and human-made music don't have to compete in the same category. In the Olympics, for example, ladies and gentlemen don't compete against each other in the same events. Nobody considers that a statement about one group being inferior to the other. It's an acknowledgement that they're operating under different conditions and mixing them together would make the competition meaningless. Why can't music work the same way? Have a human chart, have an AI assisted chart, have a mixed category if you want. Let listeners choose. The people who want to hear AI music can listen to it. The people who want human music get a space where they know that's what they're hearing. The current approach though, banning AI from the existing charts, frames the conversation as AI or human. What if it's actually AI and human in their respective spaces?

Because here's what gets lost when slop dominates the conversation. AI, used thoughtfully, is an extraordinary equalizer. Think about the finance analyst who can spot a market correction forming from three standard deviations of movement in a bond yield curve, but cannot write a coherent paragraph about what they found. Every industry has people like this whose expertise vastly outstrips their ability to communicate it. Brilliant strategists who cannot draft a persuasive email. Researchers who understand their field at a level most people never will, but whose papers are unreadable because clear writing was never their gift. Or the person [snorts] who knows exactly what's wrong in their marriage, who has the emotional intelligence to understand the dynamic, but who freezes every time they try to say it out loud. How many conversations have ended in frustration, not because the thinking was wrong, but because the words just kind of came out a bit wrong? Or the software engineer trying to explain to their family at Thanksgiving what they actually do for a living, only to end up sounding like they're describing a money laundering scheme. How many genuinely brilliant ideas have died in somebody's head because the person who had them simply wasn't built to put them into words properly.

There are people propelling the world forward right now in labs, in trading floors, in engineering departments, who are functionally not that able to communicate what they know to the people who need to hear it. Nobody failed them. They think in systems, in patterns, in numbers, and the translation into natural language has been a bottleneck their entire careers. It's okay. AI doesn't replace the intelligence in any of these cases. It just translates it. It takes the signal, I'm talking the analysis, the insight, the feeling, and helps it reach the people who need to actually hear it. That is not slop. That is a bridge between what somebody knows and what they can express.

The position of this channel has always been the same, and I want to make it extremely clear again. It is humans with AI, not humans or AI. Yes, I mean an XOR there, but I'm trying to make this video accessible. Every time somebody frames this as a binary, pro AI versus anti AI, human versus machine, authentic versus synthetic, they're making the conversation empty. The binary is analytically useless. It has always been analytically useless. The interesting questions here are about degree, intent, judgment, and what the human is actually contributing to the process.

And the responsibility for getting this right does not sit with any single company or any single government or any single YouTube creator with a PhD and strong opinions. It sits with all of us. The AI companies need to stop promising mass displacement as a feature. This is just not okay. The platforms need to fix the incentive structures. They make garbage profitable. Cybersecurity, content authentication, platform design, every element of the digital ecosystem needs to adapt in parallel to the reality that bad actors can and will exploit whatever tools exist. You cannot build the engine and hope that somebody else builds the brakes.

But equally, the rest of us need to resist the urge to flatten a genuinely complex technology into a single dismissive word. Not everything AI touches is slop. Not everything a human produces is sacred. Humans have been generating low-quality, high-volume, attention-seeking garbage since long before anyone trained a neural network on the internet. The internet was not some pristine garden of human creativity before 2023. Let's just be honest with ourselves.

But here's a question I haven't seen anyone ask. If 57% of the internet is now bots and the remaining 43% is humans and some of those humans, sure, are using AI to help them think, write, create, and communicate, then who exactly is the internet for anymore? I mean, at this point, the internet is just two autonomous server farms aggressively serving targeted ads to each other while a few bewildered humans kind of stand in the corner trying to find a recipe that isn't preceded by an eight-page AI-generated essay about the author's childhood, no? The business model that financed the open web for 30 years was built on one assumption, that the traffic was human and human attention could be sold to advertisers. That assumption is slowly dying, and nobody has replaced it with anything yet.

Maybe the era of the undifferentiated feed, where everything competes for the same eyeballs in the same stream, is simply over. The future might be smaller, curated spaces where provenance matters. It might be disclosure labels, or content credentials, or something nobody has invented yet. I don't know, but the current conversation, where the entire discourse is compressed into a single word that means everything and nothing at the same time, that is not going to get us there. Slop is a problem. The flood of slop is of course a problem, but the word itself has become its own kind of slop. A low-effort, high-volume label applied indiscriminately until it's lost whatever precision it once had.

But if you think the slop flooding your feed is bad, wait until you see what happens when the company behind the most prominent AI in the world cannot stop its own failures from compounding. I made a video about OpenAI and Sam Altman that picks up exactly where this one leaves off. Thank you so much for watching. I'll see you in the next one.