Bertuccio

Collecting stories about AI businesses

[Bertuccio @bertuccioer]

A guy took his brother's channel from $0 to $214 a day in under a month with 5 Claude strategies. He's personally made over $10 million on YouTube.

No equipment. No experience. No special skills.
Five strategies and five Claude prompts he hands over to run each one yourself.

Strategy one is the game itself. YouTube has near-zero cost to start, which means heavy competition, so you post consistently, at least once a week. He uses the pottery-class study: the group graded on quantity made better pots than the group chasing one perfect one. You can't read your way to good. You roll the dice until one video pops.

Strategy two turns a 100-sided die into a 6-sided one. Three levers raise your odds: the idea, the packaging, and the value you deliver. Most people burn three months on one immaculate video, get 50 views, and quit. Wrong move. Roll often, roll smarter.

The big one is monetization. He's made $1.5 million from AdSense and still calls AdSense hard mode. Boogie2988 has 4 million subscribers and clears $3,700 a month. His client Carla has 1,500 subscribers, 100 to 200 views a video, and closed a six-figure deal. Sean went from $30,000 to over $500,000 a month, same channel, same views, better monetization. Start with one affiliate link in your next video and ship it this week.

Subscriber count means nothing. Monetization is everything.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[Bertuccio @bertuccioer]

A guy spent $70 on one faceless AI video and it paid back $65,673. One upload, 30 minutes of his time.

No filming. No face. No daily grind. He used AI to write and build the content, paid an editor $70 to drop in the background footage, and uploaded once. That was the entire investment.

The single video did the rest. 9.7 million views. 33,000 new subscribers. $65,673 back on a $70 spend. He kept staring at the numbers saying it couldn't be real.

He's not in the video. No voice he recorded, no camera, no editing himself. AI made the content, an editor stitched it, and one upload turned $70 into a return no normal business comes near.

He pulled up the analytics again like he still didn't believe it.

$70 in, $65,673 out, off one faceless video and 30 minutes.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[Bertuccio @bertuccioer]

A guy opened his YouTube analytics on camera. $300,000 from 53 million views, off 10 minutes of work a day.

No filming. No editing. No team.
The channel runs on faceless videos he barely touches, and the whole operation fits into 10 minutes each morning. He wasn't busy. He was just reading the numbers off the screen.

The dashboard told the story. 53 million views. 2.4 million watch hours. 740,000 subscribers. Over $300,000 paid out. He scrolled through it slowly, like it was someone else's account.

He's not on screen in a single video. No face, no voice he had to record himself, no daily upload grind. Ten minutes a day built a 740,000-subscriber channel that prints in the background.

He closed the laptop and shrugged.

$300,000 and 740,000 subscribers, off 10 minutes a day and a face nobody's ever seen.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[Bertuccio @bertuccioer]

A faceless YouTube channel made $61,000 last month with 12 videos of bad MS Paint drawings.
First upload went live one month ago. 130,000 subscribers. 14 million views. The top video sits at 7 million.
The creator in this video rebuilt the entire format in 20 minutes.
The stack:

Claude Code running the Higgsfield skill to draw every frame
TurboScribe to transcribe the voiceover and pull exact timestamps
One image per timestamp, 100+ generated per video

The part nobody says out loud: the drawings swap every 2 seconds. That pace is the retention engine, not the art.
Claude then downloads each image and renames it to its timestamp, so editing turns into drag, drop, done. No syncing by ear, no listening back.
The skill is free. This window closes the second everyone finds it.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[Bertuccio @bertuccioer]

$62,000 in his first 30 days. $2,000 a day. 5 minutes per video.

A creator who has run YouTube since 2011 broke down the new faceless AI channel model.

His new channel uses AI scripts, AI voice, AI visuals. He reports $62,000 in ad revenue the first month and close to $2,000 a day since.

What changed: YouTube now runs Gemini inside the algorithm and matches brand-new uploads to the right audience within days. He points to a channel with 7 uploads pulling 800,000 views on its first video.

His one rule: no full automation. Wire everything through n8n with no human and you get demonetized, shadowbanned, or banned.

He picks each new niche in Claude, then asks for 10 facts no other channel covers.

Every video is land that keeps paying after you stop uploading.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[Bertuccio @bertuccioer]

$900,178 at 19, parked next to the AMG GT S it paid for

The caption is two words. 900k at 19.

He's in the driver's seat first, headphones on, showing the phone. Shopify Analytics, last 365 days. Gross sales $900,178.79. 5,611 orders. 4,569 fulfilled. 10.54 percent of buyers return.

Then he gets out and walks to the back of the car. Mercedes-AMG GT S, silver, the GT S badge and the rear wing in frame.

The math under the flex: $900k in twelve months means the store cleared more per day than most 19-year-olds make in a month.

Same store, same numbers as the clip before. The account name says the rest. ECOM_RON.

The car has a badge. So does the dashboard.

[Born to gamble @borntogambles]

https://t.co/Xj7p9F8SVi

[Bertuccio @bertuccioer]

$43,000 in 30 days, zero subscribers, never watched a single football game.

Everyone posting World Cup shorts this summer makes nothing.

The money sits in long-form faceless videos, and two algorithm changes open a window that closes when the tournament ends.

Part 1 Feed the Gemini engine. Google wired Gemini into YouTube. The AI reads your video and finds the audience itself, so a zero-subscriber channel gets matched instantly. No following required. Channels at zero last month now pull millions of views.

Part 2 Post before kickoff. YouTube needs days to learn who a channel is for. Start on opening day and you're already late. He started weeks early so match-day videos ranked before the whistle.

Part 3 Target the casual. Hardcore fans follow established channels. The untapped audience is the American watching in their backyard for the first time, googling what offside means. The US, Canada, and Mexico carry the three highest RPMs on the platform.

Half a million views at a $10 RPM is $5,000 from ads alone. Betting apps and affiliates triple it. Football is evergreen, so the channel doesn't die after six weeks.

He won't watch the matches. He'll bank the audience watching them.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[Bertuccio @bertuccioer]

He built a whiteboard animation video in 5 minutes for $0 using 3 AI tools.

History and fact channels pull millions of views with these. Most creators assume it takes animation skills and a budget. It takes neither.

Prompt 1 into ChatGPT returns 20 whiteboard ideas. Pick one or feed your own.

Prompt 2 turns that idea into full scene prompts plus a matching voiceover script.

AI draws the hand wrong half the time. A separate hand-correction prompt rewrites the instructions so the strokes look hand-drawn.

Each scene prompt goes into Google Flow. Set video, ingredients, model, duration. 4 scenes built the example.

ChatGPT writes the final script. ElevenLabs voices it.

Voiceover on the timeline first, then the 4 scenes trimmed to match. Scale the footage to bury any watermark.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[Bertuccio @bertuccioer]

He Bet He Could Build a $20,000 YouTube Channel Faster Than Rolex Could Fix His Watch

A creator smashed his Rolex. The repair quote came back: 10 weeks. He thought that was absurd, so he made a bet against the watchmaker.

He would build a brand-new faceless YouTube channel from zero and earn enough to buy a replacement before Rolex handed back the old one.

He picked the celebrity niche, dug up a dead channel he started at 13 for the aged-domain trust, and slapped a caricature logo on it. First video went up on a channel with 1 subscriber. That subscriber was him.

It woke up to 60,000 views. 17,000 hours of watch time overnight, on a first upload, when monetization needs 4,000. The only thing missing was subscribers: 158 against the 1,000 required.

Two weeks in, the second video cleared the threshold.

Application accepted, reviewed in 4 hours, monetized.

Then the dashboard showed a 5-cent RPM. 1.3 million views had paid out $193. He stopped checking for 4 days, convinced he was done. He came back to a glitch correcting itself: $871 in one day, then $829, RPM climbing to $12.

The early-channel boost faded and later videos flopped, the way every new channel eventually gets throttled. He kept uploading anyway. 14 videos across 6 and a half weeks.

Zero to 3,800 subscribers. Over $21,000 gross. After paying the writer and the editor, $13,560 clear.

He beat the watchmaker by 3 weeks and bought a Rolex with the difference.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[Bertuccio @bertuccioer]

A faceless stickman channel made its first $9,400 in 1 months with zero animation, zero paid tools, zero drawing skill.

14 uploads. 137,000 subscribers. 14,500,000 views. Just still stick figures stitched together, 5-to-15-minute videos, first one live 8 weeks ago.

The whole pipeline runs on free plans and 1 prompt.
Claude's free plan returns 5 viral topics, then a full script, then a download.

The rule that breaks most creators: voice over first, scene second. Generate scenes first and the audio collides with them, and viewers click away before they can explain why.

ElevenLabs free plan makes the voice. A transcriber finds every pause to the frame, so the audio tells you where each of the 103 cuts lands. No guessing. Claude writes an image prompt per timestamp. Google Flow with Nano Banana 2 renders them. A free Chrome extension batches all 103 and auto-downloads in the background. 102 land clean. 1 fails. Fix it by hand.

Sync by filename timestamp, preview for rhythm, export.

$0 setup. 103 scenes. 1 published video built by someone who can't draw.

The channel crossed another million views while you read this.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[Bertuccio @bertuccioer]

$1,300 in 6 minutes. One roofing company that didn't know it needed a website yet.

Here is the play.

Open Google Maps. Search a trade in any city.

Roofers, plumbers, landscapers. Look for the ones with little to no online presence. Closed listings, no website button, three reviews.

Copy the business info straight off the map. Drop it into an AI site builder. Six minutes later you have a clean, finished site for a company that has nothing.
Now the move: do not publish it.

Call the owner. "I built your website. Want it for thirteen hundred?"

If they say no, you lost 6 minutes. If they say yes, you just made $1,300 on work that was done before you dialed.

The whole thing fit inside one coffee break. You sold something that existed before the owner knew it did.

[Moysei @0xMoysei]

https://t.co/VXd8WiocaV

[Bertuccio @bertuccioer]

He hit $90,000 in a single month and has never sold an AI website. Stop selling AI websites in 2026.

Every YouTube channel sells the same play: build AI websites, collect recurring revenue, win.

He ran the math and built his agency past $90,000 in one month without selling a single one.

The website is a one-time product. You charge $200 to $500, maybe $3,000 if you land an enterprise client. Delivery ends, the client leaves, you start next month at zero.

To clear $50,000 a month at $500 a site, you need 100 new clients. One closed every working day, forever.

A guy in India builds the same site for $200. With everyone using the same tools, $5,000 for a website nobody can attribute revenue to is a dead pitch.

He sells filled calendars instead. Ads plus booking bot plus SMS follow-up plus automated reports plus Google reviews. The client's only job is showing up to appointments. He charges $3,000 to $5,000 a month, they stay 6 to 12 months, one client is worth $18,000 to $60,000.

Three new clients a month at $4,000 each puts $12,000 in the owner's pocket. They pay him $5,000, net $7,000. Cancel and that $7,000 disappears. Nobody cancels.

To hit $10,000 a month he needs four clients. The website model needs 20.

Roofers, dentists, landscapers all want the same thing, and it isn't a website.

They don't pay $5,000 for AI. They pay because you hand them $12,000.

[Born to gamble @borntogambles]

https://t.co/Xj7p9F8SVi

[Bertuccio @bertuccioer]

$130,000 in 37 days. The product was a PDF anyone could have Googled for free.

A relationship-niche creator typed "situationship exit strategy" into a trend tool. Opportunity score: 95 out of 100. Rising search, low competition.

He pasted the title into Claude. 15 minutes later he had a complete designed guide. Dropped it into Canva, listed it on Shopify for $9 to $27.

No ad budget. He paid one YouTuber with 28,000 subscribers $500 for a single mention. The video pulled 500,000 views.

Conversion rate: 4.19%. May 1 to June 7, the dashboard read $130,000.

The buyers had already searched the answer. They paid $19 to stop searching.

[Born to gamble @borntogambles]

https://t.co/Xj7p9F8SVi

[Bertuccio @bertuccioer]

@0xMoysei i'm trying!

[Bertuccio @bertuccioer]

$1,400 in 30 minutes. One phone call, one website, zero code.

Here is the play.

Open Google Maps. Find a local business with no website or an ugly one. Plumbers, roofers, dentists still running on nothing but a phone number.

Copy their info. Drop it into an AI site builder. Ten minutes later you have a clean, finished site.

Now the move: do not publish it.

Call the owner. "I built your website. Want it for fourteen hundred?"

If they say no, you lost 10 minutes. If they say yes, you just made $1,400 on work that was done before you dialed.

The whole thing took 30 minutes. You sold something that existed before the owner knew it did.

[Moysei @0xMoysei]

https://t.co/VXd8WiocaV

[Bertuccio @bertuccioer]

I burned $3,000 in API tokens testing PewDiePie's new AI project so you do not have to.

Save your money. Save your 3 days. It is hype wearing a hoodie.

Here is what actually happened.

Everyone online is screaming about how revolutionary it is.
A cookbook. A local agent. Runs scripts on your own machine. No subscription. Your computer becomes the model. It reads your files, sends your emails, connects to your tools. Sounds like the future.

So I did the thing nobody bragging about it actually did.

I ran it. For real. For 3 full days.

And it fell apart in my hands.

Setup that fights you at every step.
Documentation that assumes you already know the answer.
An agent that confidently does the wrong thing, then does it again.
Tasks that a free tool finishes in one click took me an hour of babysitting.

By day 2 I was not building anything.
I was debugging a toy.

By day 3 I had a $3,000 token bill and nothing I would ship to a single real user.

So why is everyone calling it genius?

One reason.

100 million subscribers.

When a name that big drops a project, people do not test it.
They retweet it.
They tutorial it.
They sell courses about it before they have run a single command.

The hype is not measuring the product.
It is measuring the follower count.

A small dev ships the same thing and it dies in silence.
A famous one ships it and it trends for a week.

Same code. Different face. Wildly different reaction.

I am not telling you fame is fake.
I am telling you fame is not a code review.

Do not download something because the person who made it is huge.
Download it because it survives 3 days of you actually trying to break it.

Mine did not.

Your hype is somebody else's marketing budget.

[Neuro Club @NeuroClubAi]

https://t.co/ZHBOEtlTf1

[Bertuccio @bertuccioer]

He made a business with the help of AI and FIFA, generated money. This farm brought him $400,000 a month.

[Bertuccio @bertuccioer]

A 14 year old just built a working drone without knowing a single thing about electronics.

he didn't study. he didn't solder a test board first. he typed one sentence into an AI and it handed him the entire machine.

this might be the most dangerous tool I have seen all year.

it is Claude Code, but for the physical world.

you type what you want to build. a drone. a boat. a greenhouse. a solar tracker that follows the sun on its own.

and the AI spits out everything.

the full wiring diagram, every wire to every pin.
the exact parts list, every component down to the resistor.
a step by step assembly guide a literal beginner can follow.

for Arduino. for Raspberry Pi. for an FPV racing drone you fly the same day.

100% free.

think about what just happened here. electronics used to take years. a degree. a lab. a mentor. a graveyard of fried boards before anything worked.

now it takes one prompt and an afternoon.

the barrier didn't get lower. it got deleted.

and this is the pattern nobody is ready for. AI is not just writing your code anymore. it is designing the thing in your hands. the object on your desk. the machine in the air.

the people who win the next ten years are not the ones who memorized how it works.

they are the ones who learned to just ask.

this is exactly what people are waking up to inside @NeuroClubAi. not watching the future happen. building it before everyone else catches on.

stop waiting for permission to build. the tools are already free. the only thing missing is you typing the first sentence.

[Neuro Club @NeuroClubAi]

https://t.co/ZHBOEtlTf1

[Bertuccio @bertuccioer]

This is what a money farm looks like in 2026, and there is not a single human working it.

people used to hire hands to work the fields. now the machines do the labor and write the content too. he runs UGC production and account growth completely automated, and it pays him around 70 thousand dollars a month.

here is what the farm actually is.

instead of crops, he grows accounts. instead of workers, he runs software. one system spins up profiles that look completely human, ages them until the algorithm trusts them, and then sets them loose to produce content on their own. user generated content that was never touched by a user. posts, captions, comments, all of it manufactured and scheduled while he sleeps.

then it scales the way a real farm scales. more accounts, more output, more reach, with zero extra hands. each profile grows its following, builds its history, and starts pulling real attention. and that attention is the harvest. he sells it as UGC production and account growth to brands and clients who think they are buying real audiences and real creators.

then the part that should land.

the old farm needed land, workers, and seasons. this one needs a dashboard and patience. he plants accounts, the machine grows them, and he sells the yield for 70 thousand a month while the fields run themselves.

the work did not disappear. it just stopped needing people. and the ones who figure out what to farm next are the ones who eat.

[Boney R. @boney2r]

High school kids are banking 15,000,000 dollars a year from one app they built in a weekend

They open the App Store.
Pick the top mail tracker.
Screenshot the app, every feature, every review.

Drop the images into Claude.

Ask who the real customers are.
Ask what problem it actually solves.
Ask where the reviews reveal the gap.

Then the real move.

Tell Claude to add AI that removes the friction.
Voice logging instead of typing.
Say two eggs and toast and it logs everything.

That single idea became the product.

They shipped it with zero design skills.
Zero code.
Just dangerous permissions and Claude.

This is the new app factory.
One screenshot at a time.