CTRL+WATCH

AN IN-DEPTH EXAMINATION OF THE SYSTEM THAT DECIDES WHAT 2.7 BILLION PEOPLE WATCH

The most powerful editor in the history of media has no face, no byline, and no editorial meetings. It processes over 500 hours of uploaded video every minute. It decides which creators eat and which starve. It has reshaped human attention on a planetary scale. And almost nobody — including the people who built it — can tell you exactly how it works.

We're talking about the YouTube recommendation algorithm: the system responsible for an estimated 70% of all time spent watching videos on the platform. Not the search function. Not your subscription feed. The recommendation engine — the thing that puts videos in front of you that you didn't ask for but can't stop clicking. It's the most consequential piece of software in the entertainment industry, and we've spent the last two months trying to understand it.

What follows is not a conspiracy theory. It's not a hit piece on YouTube. It's an attempt to describe, as clearly and honestly as possible, what the algorithm actually does, who benefits, who suffers, and what it means for the future of creative work on the platform we love.

CHAPTER 1: WHAT THE ALGORITHM ACTUALLY IS

First, let's kill a myth: there is no single "YouTube algorithm." The term is shorthand for a sprawling network of machine learning models that collectively decide what to recommend, what to surface in search, what to suggest in the sidebar, and what to feature on the homepage. These systems evolved from a 2016 Google Brain paper that introduced deep neural networks to the recommendation pipeline, replacing the older view-count-and-star-rating system that YouTube relied on in its early years.

The current system operates in two stages. The first is candidate generation: from hundreds of millions of videos, the system narrows the pool to a few hundred that might interest a given user. The second is ranking: those candidates are scored and ordered based on predicted engagement — primarily watch time, but also click-through rate, likes, shares, and a metric YouTube calls "satisfaction" that incorporates survey data.

Think of it like this: candidate generation is the bouncer deciding who gets into the club. Ranking is the DJ deciding who gets the spotlight. Both are operating on probabilistic models trained on billions of data points, and both are optimizing for a single overriding goal: keep you watching.

CHAPTER 2: THE THUMBNAIL ECONOMY

The algorithm's reliance on click-through rate has created what we call the Thumbnail Economy — an entire secondary industry built around the science of making people click. Professional thumbnail designers now command fees of $500-$5,000 per image. A/B testing platforms let creators test six different thumbnails before a video goes live. The "surprised face" thumbnail has become so ubiquitous that it's essentially the visual language of modern YouTube.

This isn't inherently bad. A good thumbnail is an act of communication. But the optimization pressure is relentless. Creators describe a treadmill effect: each video needs to perform as well as the last, or the algorithm reduces your reach. You're not competing with other creators in your niche. You're competing with every piece of content on the platform, from NASA launches to cat videos to geopolitical commentary. And the algorithm doesn't care about your artistic vision. It cares about whether someone clicked, and whether they stayed.

CHAPTER 3: THE CREATOR'S DILEMMA

Every serious YouTube creator faces the same impossible choice: do you make the video you want to make, or the video the algorithm wants you to make? The answer, for anyone who wants to pay their rent, is usually some compromise between the two.

The algorithm rewards several specific behaviors. Longer videos (eight minutes or more to qualify for mid-roll ads). Consistent upload schedules. High-retention openings that hook viewers in the first 30 seconds. Titles and thumbnails optimized for curiosity gaps. Topics that match trending search queries. These aren't creative decisions — they're engineering requirements.

The result is a kind of natural selection. Channels that adapt survive. Channels that don't, decline. And the traits being selected for aren't necessarily the same traits that make content good. The algorithm can't measure insight, beauty, truth, or originality. It can measure whether someone watched for eight minutes instead of four. The gap between those two things is where a lot of YouTube's creative soul gets lost.

Linus Sebastian himself admitted in late 2025 that LTT was "on the struggle bus" after a sudden viewership drop, noting that the first few minutes of viewer engagement "basically makes or breaks a video." When a channel with 16 million subscribers and daily uploads can be brought low by an algorithmic shift, what hope does a channel with 16 thousand have?

CHAPTER 4: THE ALGORITHM'S VICTIMS

For every Mark Rober success story, there are thousands of channels that were thriving one month and invisible the next. The algorithm giveth, and the algorithm taketh away, and it never sends a memo explaining why.

Educational channels are particularly vulnerable. A well-researched, carefully produced 20-minute explainer about Byzantine architecture will lose the recommendation battle to a 10-minute video titled "I Ate the World's Spiciest Chip" every single time. Not because the audience doesn't want education — the success of channels like Kurzgesagt and 3Blue1Brown proves they do — but because the algorithm's optimization function doesn't have a category for "intellectual value." It has a category for "predicted watch time," and the spicy chip video wins that race by a mile.

The pattern is consistent: the algorithm amplifies content that provokes strong, immediate emotional responses. Outrage, shock, curiosity, fear, FOMO. These emotions drive clicks and watch time. Nuance, complexity, and measured analysis drive thoughtful engagement — which the algorithm can't distinguish from boredom.

CHAPTER 5: THE ALGORITHM'S WINNERS

Not everyone is fighting the algorithm. Some creators have learned to ride it like a wave. MrBeast is the obvious case study — a creator who reverse-engineers the recommendation system with the precision of a Wall Street quant. Every title, thumbnail, and video structure is designed to maximize algorithmic distribution. It works. He has over 300 million subscribers and his videos routinely cross 100 million views.

But the more interesting case studies are the creators who've found the sweet spot between algorithmic optimization and genuine quality. Mark Rober uploads once a month and gets 30-50 million views per video because each one is so inherently shareable that the algorithm can't ignore it. MKBHD maintains such consistent quality that his audience retention numbers are abnormally high, which the algorithm rewards with premium recommendation placement. Conan O'Brien's Team Coco channel benefits from what you might call "legacy authority" — years of consistent engagement data that give the algorithm confidence in recommending the content.

The lesson, if there is one, is that the algorithm isn't anti-quality. It's quality-blind. It can't see quality. It can only see the metrics that sometimes correlate with quality. Creators who understand this distinction have an enormous advantage over those who don't.

CHAPTER 6: THE SATISFACTION PROBLEM

YouTube knows it has a problem. In recent years, the company has made a concerted effort to move beyond raw engagement metrics toward what it calls "viewer satisfaction." This includes post-viewing surveys, the introduction of the "not interested" button, and changes to the recommendation system designed to reduce the promotion of "borderline content" — videos that don't violate policies but aren't exactly what you'd call wholesome.

The shift is real but incomplete. Satisfaction surveys are inherently limited — people don't always know what's good for them, and the gap between "what I watched for an hour" and "what I'm glad I watched for an hour" is enormous. YouTube is trying to solve one of the hardest problems in AI: building a system that optimizes not just for engagement, but for genuine value. Given that humans have been arguing about the definition of "genuine value" since Plato, we wouldn't hold our breath.

CHAPTER 7: OUR VERDICT

The YouTube algorithm is not evil. It's not good. It's a mirror — a mirror built from engagement data that reflects humanity's preferences back at us, amplified and distorted. When we complain about the algorithm promoting shallow content, we are, to some degree, complaining about ourselves. The algorithm didn't create our appetite for outrage, spectacle, and novelty. It just learned to serve it more efficiently than any system in history.

But a mirror that distorts is not a neutral object. The algorithm doesn't just reflect preferences — it shapes them. It creates feedback loops where success breeds more of the same, where experimentation is punished, where the safest creative choice is also the most rewarded one. It has flattened the aesthetic of YouTube into a recognizable formula: the surprised face, the bold title, the eight-minute runtime, the mid-roll ad at 4:02.

The best creators on YouTube are the ones who understand the algorithm well enough to work with it, but have enough artistic conviction to work against it when it matters. They know the rules, and they know when to break them. That's always been the definition of great art — and the fact that it applies to YouTube is either depressing or inspiring, depending on how you look at it.

We built a magazine. You're reading it. The algorithm can sort itself out.

END OF INVESTIGATION

We showed Andy Warhol a selection of top influencer content from 2025 — beauty tutorials, unboxing videos, "day in my life" vlogs, brand deals, and the general machinery of the creator economy. He was seated in what appeared to be a very deliberate arrangement of indifference. He had been watching for approximately ninety minutes. He had not blinked enough.

We showed Kurt Cobain the current state of music on YouTube: reaction channels, lyric videos with 2 billion views, algorithmically optimised lo-fi playlists, bedroom pop artists with no label deals and millions of streams, and the comment section under "Smells Like Teen Spirit" (currently at 1.8 billion views). He sat cross-legged on the floor. He did not sit in the chair provided. He looked tired in the way that only people who feel everything look tired.

9.2M SUBSCRIBERS • EST. 2010 • COMEDY / PODCAST / ENTERTAINMENT

There is a theory in entertainment that the best late night hosts don't adapt to new formats — they simply wait until the format bends to accommodate them. Conan O'Brien is the proof. After a 28-year late night television career that ended not with a whimper but with a controlled demolition of the genre itself, Conan has built something on YouTube that is, frankly, better than anything he did on network or cable television.

Team Coco on YouTube started as a clip dump. Literally just segments from the TBS show, chopped up and uploaded. But something happened: the clips became more popular than the show. The Clueless Gamer segments. The remote bits with Jordan Schlansky. The Conan Without Borders travel specials. The algorithm loved them because they were self-contained stories with enormous replay value, and audiences loved them because Conan is one of the funniest human beings alive.

In May 2025, Team Coco made the leap to full-length video podcast episodes of "Conan O'Brien Needs a Friend" on YouTube. This was the missing piece. The podcast — already ranked in Edison Research's Top 20 — translates beautifully to video. Conan's physical comedy, his ridiculous tangents with Sona Movsesian and Matt Gourley, the genuine warmth of his celebrity interviews: it all plays better when you can see it. The man hosted the Oscars in 2025, won the Mark Twain Prize, launched a second season of his HBO travel show, AND started dropping weekly video podcasts on YouTube. He's 62 years old and outworking creators half his age.

The weak spots are minor but real. The channel's identity was muddled for years — too many clip formats, inconsistent upload patterns, a lingering sense that YouTube was the B-side rather than the main event. And while Team Coco has 9+ million subscribers, the comment section engagement doesn't match that of native YouTube channels. This is a traditional media audience that migrated, not a community that was born on the platform. You can feel the difference.

6.2M SUBSCRIBERS • EST. 2021 • SCIENCE / HEALTH / EDUCATION

Andrew Huberman is the most influential and most problematic science communicator on YouTube. Both things are true simultaneously, and you need to hold them in your head at the same time to understand what his channel actually is.

Let's start with what works. Huberman Lab episodes are, at their best, remarkable feats of science communication. The deep dives into neuroscience, sleep, and vision — his actual areas of academic expertise — are genuinely excellent. He explains complex mechanisms in language that an intelligent layperson can follow without feeling patronized. The production quality is clean, the pacing is deliberate, and there's an authority in his delivery that comes from being a working Stanford neuroscientist who publishes real papers in real journals. When Huberman stays in his lane, he's doing something no one else on the platform can match.

The problem is that Huberman almost never stays in his lane. The podcast has metastasized from neuroscience explainers into a sprawling wellness empire that covers everything from gut health to testosterone optimization to supplement protocols — topics well outside his area of research expertise. Scientists like Jonathan Jarry have criticized his promotion of poorly regulated dietary supplements. Cancer biologist Joseph Zundell has noted his tendency to extrapolate animal studies to human application without scientific justification. New York Magazine described his podcast as a place where Huberman "posits certainty where there is ambiguity."

And then there's the TRT revelation. In early 2026, Huberman confirmed he's been on testosterone replacement therapy since age 45 — while simultaneously promoting (and sponsoring) supplements designed to naturally boost testosterone. The hypocrisy isn't subtle. His subreddit called him a "scam artist" and a "charlatan," and while that's harsh, the credibility hit is real and deserved. When your monetization model depends on people buying supplements because they trust your scientific authority, and you've been quietly using pharmaceutical-grade testosterone the entire time, you have a transparency problem that no amount of morning sunlight is going to fix.

The community score suffers most. Huberman Lab has spawned a "protocol bro" subculture that treats his podcast episodes like religious texts — dogmatic adherence to morning routines, cold exposure timings, and supplement stacks that are presented with far more confidence than the underlying science warrants. This isn't entirely Huberman's fault, but he's done nothing to discourage it, and his CBS News appointment suggests he's happy to expand his influence rather than sharpen his focus.

4.3M SUBSCRIBERS • EST. 2006 • COMEDY / PODCAST

Theo Von is the closest thing YouTube has to a jazz musician. You never quite know where a sentence is going to land, and when it does, it lands somewhere you couldn't have predicted but somehow feel like you've been before. He tells stories about growing up in Covington, Louisiana, that sound like dispatches from a dimension adjacent to our own — familiar enough to be funny, strange enough to be genuinely revelatory.

The channel is built around "This Past Weekend," his long-running podcast that has quietly become one of the most compelling interview shows on YouTube. The format is simple: Theo talks to people. That's it. But the execution is extraordinary. His interviews with comedians, fighters, and cultural figures have a loose, improvisational energy that makes guests say things they wouldn't say anywhere else. His solo episodes — where he riffs on topics sent in by listeners — are masterclasses in comedic storytelling. A twenty-minute Theo Von monologue about something that happened at a gas station in 1997 will make you laugh, then think, then feel something you didn't expect to feel.

Production value is minimal and that's the point. The set is simple. The editing is basic. The thumbnails won't win any awards. But Theo's genius is in the language itself — his metaphors come from a place that no writer's room could manufacture. His descriptions of childhood, poverty, family dysfunction, and the American South are simultaneously hilarious and deeply compassionate. He is a profoundly empathetic comedian who disguises his empathy as absurdism, and the result is content that hits emotional frequencies most podcasters can't even hear.

The clip channels extend his reach enormously. Theo Von Clips has 1.3 million subscribers and serves as a discovery engine for the main show. The comment sections across both channels are among the best on YouTube — genuinely funny, engaged, and weirdly wholesome for a comedy community. His audience gets it. They're not there for takes or discourse or drama. They're there because Theo Von makes them feel less alone, and he does it while making them laugh so hard they have to pull over.

70M+ SUBSCRIBERS • EST. 2011 • SCIENCE / ENGINEERING / EDUCATION

Mark Rober has solved YouTube. Not in the cynical, algorithm-gaming sense. In the way that a NASA engineer solves problems: by understanding the system, identifying the constraints, and building something that works within them while being genuinely remarkable.

Every Mark Rober video follows the same formula, and it works every single time: take a scientific or engineering concept, build something insane to demonstrate it, film the process with impeccable production value, explain the science in language a ten-year-old can follow, and end with a spectacular payoff that makes you want to share the video with everyone you know. Glitter bombs for porch pirates. The world's largest Nerf gun. A squirrel obstacle course that became a cultural phenomenon. The roller coaster he built in his lab. Each video is a self-contained engineering marvel, and each one functions as a perfectly designed package for algorithmic distribution.

The numbers are staggering: 70+ million subscribers, 13.7 billion total views, and a catalog of just over 220 videos. Do the math. That's roughly 62 million views per video on average. Most channels would sell a kidney for those numbers on their best day. Rober gets them as a baseline, and he does it while uploading maybe twelve times a year.

The 2025 expansion was extraordinary. A Netflix deal for a greatest-hits compilation and a new competition series. The Team Water campaign with MrBeast raising $40 million for WaterAid. CrunchLabs evolving from subscription boxes into a full media company with a school curriculum division. A TIME100 Creators listing. A Sesame Street Christmas special. Mark Rober is no longer a YouTuber who does science. He's a science media empire that happens to publish on YouTube.

The consistency score is the only thing keeping him from ESSENTIAL territory. Twelve uploads a year is remarkable per-video output, but it means months of silence between uploads. In a platform economy that rewards frequency, Rober's strategy is an anomaly — proof that sheer quality can override the algorithm's preference for quantity, but also an approach that only works at his scale. His CrunchLabs integrations, while tasteful by YouTube standards, have a faint corporate sheen that occasionally undercuts the DIY wonder. And the community, while massive, skews young and family-friendly in a way that limits the depth of engagement you see on more niche channels.

This is the tech YouTube fight everyone's been waiting for. On one side: Marques Brownlee, MKBHD, the man who made tech reviews look like cinema. Twenty million subscribers. Every frame a wallpaper. Reviews so visually gorgeous that you sometimes forget to listen to the actual opinion. On the other: Linus Sebastian, Linus Tech Tips, the relentless content machine. Sixteen million subscribers on the main channel alone, with a sprawling media empire that publishes daily across multiple channels with the energy of a caffeinated lab that never sleeps.

These are the two biggest tech channels on YouTube by influence (not raw subscriber count — that crown goes to other channels), and they represent fundamentally different philosophies about what a tech review channel should be. MKBHD is the auteur. LTT is the newsroom. MKBHD values restraint, craft, and visual storytelling. LTT values comprehensiveness, entertainment, and sheer volume. Both approaches have produced extraordinary results. Both have significant flaws.

THE CASE FOR MKBHD

Marques Brownlee has done something genuinely rare: he's made tech reviews into an art form. The camera work, the lighting, the studio design, the colour grading — every element is considered with a precision that borders on obsessive. When MKBHD reviews a phone, you're not just getting an opinion on the camera and battery life. You're getting a cinematic experience that communicates quality through every frame. He's also an outlier in terms of editorial independence — no paid reviews, transparent about sponsorships, and willing to criticize products from companies that advertise with him. In a space rife with conflicts of interest, that matters. His expansion into Auto Focus (1M+ subscribers), the Waveform podcast (438K subscribers), and The Studio shows ambition matched by quality.

The Panels wallpaper app debacle was a genuine miss — a $50/year subscription for wallpapers from a guy who reviews pricing decisions for a living was a self-own of impressive proportions. He handled the backlash reasonably (reduced pricing, eventually shut it down), but it revealed a gap between his tech critic identity and his business instincts. The channel itself, however, remains untouched by the controversy. If anything, his handling of it reinforced why people trust him: he acknowledged the mistake rather than doubling down.

THE CASE FOR LTT

Linus Tech Tips is the opposite of a one-man show. It's a media operation with 100+ employees, multiple channels, daily uploads, a weekly live show (WAN Show), and a testing lab that rivals professional publications. The breadth of coverage is staggering — LTT has reviewed, tested, or benchmarked more products than any other channel in tech YouTube history, and the sheer volume means there's an LTT video for virtually any tech question you might have. It's the Wikipedia of tech YouTube, and there's enormous value in that.

But the controversies. God, the controversies. The 2023 Gamers Nexus exposé revealed rushed reviews, the Billet Labs prototype disaster, and workplace culture allegations that forced a content pause. The 2025 feud reignited with plagiarism allegations, leaked emails, and WAN Show segments that ranged from defensive to combative. A former employee of nearly a decade left citing stagnant pay and shifting expectations. LTT's viewership took a measurable hit in late 2025, with Linus himself admitting videos were underperforming their historical benchmarks. The controversies don't erase LTT's value as an information resource, but they've fundamentally damaged the channel's credibility narrative. When your brand is "trust us, we test everything," and the testing methodology is questioned repeatedly by your peers, you have a structural problem.

THE VERDICT

This is closer than the numbers suggest, because MKBHD and LTT are optimizing for different things. If you need to know whether a specific GPU is worth buying, LTT is where you go. If you want to understand what a piece of technology means — how it feels, where it fits in the landscape, whether it matters — MKBHD is the better guide. One gives you data. The other gives you a perspective.

We're giving this to MKBHD, and the reason is simple: quality ceiling. On his best day, MKBHD produces content that is indistinguishable from professional film. On his worst day, he produces a competent tech review. LTT's best days are excellent — genuinely informative and entertaining — but their worst days have produced factual errors, ethical lapses, and content that their own peers have publicly called out. In a head-to-head, the channel with the higher floor AND the higher ceiling wins.

Editorial note: LTT's 9.2 billion total views versus MKBHD's 5.1 billion demonstrates that volume has its own power. If you're building a tech YouTube diet, you need both. But if you could only subscribe to one — which is the question Boss Fight asks — MKBHD is the one that never lets you down.

UPDATED FROM ISSUE #001 • 3 NEW ENTRIES • 3 DROPPED • 6 RE-SCORED • ARGUE ABOUT IT

The master rankings return, updated to reflect this month's reviews, re-evaluations, and the general chaos of the YouTube landscape. Three new entries join the list (Conan O'Brien, Theo Von, and Huberman Lab), which means three channels have been pushed off the bottom. Six channels have been re-scored based on significant developments since Issue #001. Position changes marked accordingly. If your favourite channel dropped, it's because someone else got better. That's how competition works.

SCORE KEY: 90+ ESSENTIAL • 80-89 EXCELLENT • 70-79 GOOD • 60-69 AVERAGE

DROPPED OFF: PewDiePie (legacy, was #48) • Baumgartner Restoration (was #49) • The Infographics Show (was #50)

THE HEART OF THE MAGAZINE • WHERE THE ALGORITHM HASN'T REACHED YET

Hidden Levels is our favourite section because it's where the algorithm hasn't had a chance to flatten everything into sameness. These are channels we found by accident, by recommendation, by stubborn refusal to stop browsing at 2am. They're small. They're weird. They're brilliant. Subscribe to them before everyone else does, and you can claim you were there first.

Mark Rober's videos look effortless. They aren't. Each one takes between two and six months to produce, involves a team of engineers, a dedicated content studio (CrunchLabs HQ), and a production process that would be familiar to anyone who's worked in Hollywood rather than YouTube.

Phase 1: The Concept (2-4 weeks). Rober and his team brainstorm ideas based on a simple test: "Can we build something that demonstrates this concept in a way that creates a visual spectacle?" If the answer is yes, and the science is sound, it moves to prototyping.

Phase 2: Engineering (4-12 weeks). This is where Rober's NASA background becomes the unfair advantage. The glitter bomb wasn't just a prank — it was an engineered system with GPS tracking, motorized spinning mechanisms, and fart spray dispensers on precision timers. The squirrel obstacle course involved weeks of testing and iteration. These are engineering projects that happen to be filmed, not filming projects that happen to involve engineering.

Phase 3: Filming (1-2 weeks). Multiple camera angles, drone footage, slow-motion captures, and carefully staged reveals. Rober's cinematography is among the best on the platform — clean, dynamic, and always in service of the story. CrunchLabs now has a Chief Content Officer (Scott Lewers, formerly of Lionsgate) managing the production pipeline.

Phase 4: The Edit (2-4 weeks). This is where the educational layer gets woven in. Rober's signature is making you learn without realising you're learning — the science explanation is never a lecture, it's embedded in the narrative of the build. Graphics, animations, and data visualizations are custom-made for each video.

Phase 5: Thumbnail and Title (Ongoing). Like every major creator, Rober A/B tests thumbnails. But his advantage is that his videos contain genuinely spectacular images — a 100-foot Nerf gun, a glitter explosion, a squirrel flying through the air — so the thumbnails are real moments, not manufactured shock faces.

SHARP BUT NEVER CRUEL • BECAUSE WE LOVE THIS PLATFORM ENOUGH TO ROAST IT

1. THE AI SLOP DELUGE

AI-generated content channels have gone from "interesting experiment" to "sewage pipe aimed directly at the recommendation feed." Faceless channels churning out AI-narrated, AI-scripted, AI-thumbnailed content about topics like "10 EXTINCT ANIMALS THAT COULD STILL BE ALIVE" are clogging the platform at industrial scale. The production cost is zero. The creative value is less than zero. YouTube's 12 million channel terminations in 2025 barely made a dent. This is the new spam, and it's winning.

2. THE THREE-HOUR PODCAST EPISODE

We need to talk about runtime. There was a time when a YouTube video over 20 minutes felt long. Now every podcast episode is a three-hour commitment that requires scheduling. Huberman Lab episodes routinely hit 2.5-3 hours. Joe Rogan regularly goes past three. Lex Fridman once did seven hours with Elon Musk. Lads: not every conversation needs to be longer than most Hollywood films. Some of us have jobs. Edit your content. Respect our time. The algorithm rewards watch time but it shouldn't reward hostage situations.

3. THE FAKE CANDID THUMBNAIL

You know the one. The creator photographed from the side, hand on chin, looking pensively into the middle distance as if they're contemplating the heat death of the universe when they're actually thinking about their CPM. Bonus points for an out-of-focus bookshelf in the background (always the same four books). The "thoughtful intellectual" thumbnail has become as formulaic as the surprised face it was supposed to replace. You're not fooling anyone.

4. THE SUPPLEMENT AD DISGUISED AS SCIENCE

The Huberman Effect has spawned an entire ecosystem of channels that present supplement advertisements as peer-reviewed research. The formula is always the same: cite one mouse study, extrapolate wildly to humans, mention the specific product name four times, and include an affiliate link in the description. This is not science communication. This is QVC with a lab coat. The audience deserves better and the FTC should be paying closer attention.

5. "I QUIT YOUTUBE" (NO YOU DIDN'T)

The genre of "I'm leaving YouTube" videos, uploaded to YouTube, promoted by the YouTube algorithm, monetised with YouTube ads, by creators who return to YouTube three weeks later, continues to thrive. We counted seventeen "I'm quitting" videos in December 2025 alone from channels with over 100K subscribers. Fourteen of them have uploaded since. The other three pivoted to Twitch, discovered it pays less, and quietly came back. If you're going to quit, quit. Don't make a content series about it.

YOUR LETTERS. YOB'S JUDGEMENT. NO APPEALS.