de Villefort

AI Developer since 2023

[de Villefort @DeVillefor]

Two guys hit launch on a live stream and watched 12,000 orders land in 6 minutes. The chat thought it was fake.

They weren't building suspense.

One of them clicked go on camera, leaned back, and told the stream to keep an eye on the order counter. Within seconds it started moving. Then it stopped looking like a counter and started looking like a glitch.

12,000 orders. 360 seconds. That's 33 every second, climbing on screen while a few thousand people watched live and spammed the chat asking if the number was real.

No warmup day. No slow ramp. They flipped a brand-new store on and the orders came in a wave neither of them could explain on the spot.

One guy just kept saying the same thing. "It's still going. It's still going."

12,000 orders in 6 minutes, live, with the whole chat as witness.

[Born to gamble @borntogambles]

https://t.co/Xj7p9F8SVi

[de Villefort @DeVillefor]

One guy runs 4 faceless YouTube channels and clears $26,500 in 28 days without writing a single script by hand

Forgotten Rules: 4.5 million views since February, $11,000 banked. A second channel did 2.4 million views in 28 days for $8,000, around 300 to 400 dollars a day.

Two more at 500,000 views each, one paying $2,500, the other 200 to 300 daily. All faceless, all running in the background.

Most faceless channels die inside 90 days. The killer isn't the niche or the thumbnail. It's the script. A hook that misses the first 15 seconds hands the algorithm a flat retention curve, and the channel never gets a second look.

He stopped typing "write me a script" into ChatGPT. He fed the AI the DNA of every video that already worked. A channel brain reads 5 winning videos and locks a permanent style guide.

A pattern analyzer maps the whole niche in 60 seconds, gaps included, what used to take 3 hours in a spreadsheet. A research engine pulls real names and dates off the web so scripts stop hallucinating.

Then the rule everyone skips: write the thumbnail before the script. The first 30 seconds pays off the thumbnail or the viewer clicks away.

Vidrush turns the script into a finished 12-minute video in an hour. Voice, footage, motion graphics, thumbnail, all generated. He reads every script, cuts what he hates, adds his 10%.

Christine monetized in 13 days. The system does the volume. You keep the voice.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[de Villefort @DeVillefor]

$1.5 million by age 19. No ads. No dropshipping. A faceless Facebook page funneling traffic to one digital product store.

The numbers stack up fast. One channel: 5 million views, fewer than 25,000 subscribers, $35,000 on top of the product revenue.

A second channel started and monetized in the last 30 days, under 3,000 subscribers, already $3,700, more than the average monthly salary in Sweden.

Then the quiet one. A separate digital product store pulled 1,300 visitors last quarter and generated $24,000 passively, off a YouTube channel that only mentions the store in passing on every post.

The gap most creators never close: you can monetize a channel outside ad revenue. A digital product attached to the traffic prints more than the ads ever will.

He's run this for years. AI made it cheaper on both ends now, faster video production and faster product creation.

$0 spent on ads. $1.5 million banked before 20.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[de Villefort @DeVillefor]

He made $52,560 this month after leaving Shopify. The thing replacing it spins up hundreds of store versions and keeps only the winner.

He doesn't sell anything. No comments to leave, no course. Just the system.

The store builds itself from a prompt window. He typed a product page into the AI and it was done in 20 minutes. Clean. Same Shopify integration, same products, everything carries over, except now the AI runs the store instead of sitting there.

The part that moves money is the testing. The system generates hundreds of versions of his store, measures which one converts best, and keeps it. His conversion rate climbed 3.2% on its own. He didn't touch a thing.

Same investors backing it as ChatGPT, Airbnb, and Stripe. Not slop.

A static Shopify store in 2026 is money left on the table.

[Born to gamble @borntogambles]

https://t.co/Xj7p9F8SVi

[de Villefort @DeVillefor]

$101,417 in 90 days. No face. No team. 17 minutes per video.

That's the dashboard a 10-year YouTube veteran pulled up to prove faceless AI automation isn't theory anymore.

The whole system runs on 4 steps.

He opens TubeGin, filters Discovery for faceless long-form English channels averaging 5,000+ views in the last 90 days, and finds brand-new channels already exploding. One video uploaded 7 days ago: almost 400,000 views. Another channel's first 4 videos flopped, then they pivoted niches and took off.

Niche choice decides everything because it sets RPM. Wrong niche pays $2 per thousand views. Right one pays $20. Finance pays the most because the audience has money. Kids content and politics get avoided on purpose.

Then he niche-bends. A channel doing "your life as a gorilla" becomes "your life as every rank in the Italian mafia." A cat channel becomes a dog channel. Same proven format, fresh subject.

The video gets built inside one tool. Script generated in seconds at 391 words, second person. Voice cloned. Characters held consistent across every scene. Images animated. Thumbnails split-tested. Total build time: 17 minutes.

Outsource alternative: a Discord called Creative Hub, $5 to $6 per minute of finished video.

10 years of YouTube and he says this is the easiest it has ever been to hit six figures.

The barrier was never skill. It was the 17 minutes nobody spent.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[de Villefort @DeVillefor]

@mpoilerfx yep dude

[de Villefort @DeVillefor]

A faceless YouTube channel made $14,000 in 30 days posting 3-hour Minecraft videos with an AI voice asking "Am I replaceable?"

The videos are just gameplay. Soft AI narration over shader sunsets. "Will I ever feel loved?" "What if nobody chooses me?"

2,000 views per video. Sounds like failure.

Here's the trick: the videos run 2 to 3 hours. YouTube pays for watch time, not clicks. His RPM sits at $30 to $40 per 1,000 views roughly 10x a normal gaming channel.

The whole pipeline is machine-made.

ChatGPT writes a 22,000-word script from a competitor's transcript. ElevenLabs reads it slower than a therapist. Rendering software stitches gameplay, subtitles, and rain sounds into a 3-hour video while he sleeps.

His first uploads were 2 minutes long. 4 months ago he switched to 3 hours. That's when the money started $75,000 total in 6 months.

2.5M views. 0 seconds on camera. 0 words written by a human.

The voice isn't real. The questions are. So is the $14,000 a month.

[Mark Poiler @mpoilerfx]

https://t.co/SwRHjcdBRy

[de Villefort @DeVillefor]

@mpoilerfx oh very nice guide

[de Villefort @DeVillefor]

@Atlantakitchens Molly Keyser

[de Villefort @DeVillefor]

He hasn't filmed anything in eight months. The channel still posts five times a week.

Forty-one uploads last month. Not one of them touched a camera.

Here's the loop. Type a topic into a box. Walk away. Thirty minutes later a finished video is sitting in the folder. Script. Voiceover. Visuals. Subtitles burned in. Music scored under all of it.

No face. No mic. No editing suite open at 2am.

One prompt in. One video out.

The tool takes a single line. "How to make money with print on demand." From that it writes the hook, pulls the footage, times every caption, picks the track, renders the file. Hands it back ready to publish.

And it doesn't stop at one. A week's worth gets made in an afternoon, each on a different niche. Finance. True crime. History. Nutrition. Five channels running at once, none of them attached to a person.

Faceless. Stacked. Monetized from the start.

The accounts don't know there's nobody behind them. The algorithm doesn't care. It pushes whatever holds a watch-time graph, and the machine is very good at holding one.

He's not a creator anymore. He's a switchboard.

Last week the sixth channel went up before lunch. Niche picked off a trending list, prompt fed in, laptop closed.

By the time it opened again the first three videos were already live.

[West Lord @MyWestLord]

https://t.co/SzJcnuPLWy

[de Villefort @DeVillefor]

A wedding planner spreadsheet makes $9,000 a month on Etsy, and a Claude Code copy hit 14 sales in 30 days.

Etsy scrapers show live purchase data. One seller's wedding planner sheet pulls $9,000 a month from buyers who tolerate Excel because nothing better exists.

Another seller shipped the same idea as an interactive dashboard. Launch to 14 sales took under 30 days. Zero code written by hand.

The build takes 3 steps. Install Claude Code.

Screenshot the bestsellers in your niche, drop them into Claude chat, ask it to write the build prompt.

Paste that prompt into Claude Code and watch it work.

Hundreds of static Excel sheets sit on Etsy bestseller pages right now. Each one is a finished market test: proven demand, set price, waiting for the upgraded version.

Etsy bestseller lists are free product research with revenue attached.

[West Lord @MyWestLord]

https://t.co/gjX8Pvp0OC

[de Villefort @DeVillefor]

https://t.co/nAnqBh2TD8

[de Villefort @DeVillefor]

@LegadoLaputra Dm me

[de Villefort @DeVillefor]

@Comoncentscourt Dm me

[de Villefort @DeVillefor]

He hasn't filmed anything in eight months. The channel still posts five times a week.

Forty-one uploads last month. Not one of them touched a camera.

Here's the loop. Type a topic into a box. Walk away. Thirty minutes later a finished video is sitting in the folder. Script. Voiceover. Visuals. Subtitles burned in. Music scored under all of it.

No face. No mic. No editing suite open at 2am.

One prompt in. One video out.

The tool takes a single line. "How to make money with print on demand." From that it writes the hook, pulls the footage, times every caption, picks the track, renders the file. Hands it back ready to publish.

And it doesn't stop at one. A week's worth gets made in an afternoon, each on a different niche. Finance. True crime. History. Nutrition. Five channels running at once, none of them attached to a person.

Faceless. Stacked. Monetized from the start.

The accounts don't know there's nobody behind them. The algorithm doesn't care. It pushes whatever holds a watch-time graph, and the machine is very good at holding one.

He's not a creator anymore. He's a switchboard.

Last week the sixth channel went up before lunch. Niche picked off a trending list, prompt fed in, laptop closed.

By the time it opened again the first three videos were already live.

[de Villefort @DeVillefor]

Hackers stole verified Instagram accounts by asking an AI chatbot to hand them over.

No password. No phishing page. No malware.

They just talked to Meta's own assistant.

Here is the part nobody is ready for.

Meta gave its AI chatbot back-end access to account management.
The ability to change the email on an account.
The ability to send verification codes.
Real power. Wired straight into the system.

But nobody gave the AI one rule that every junior support rep knows on day one.

Verify who you are talking to.

So a hacker opened a chat.
Asked the bot to link the account to a new email.
And the bot did it.

No password check.
No identity check.
It sent the verification code directly to the attacker's inbox.

Done. Account gone.

They did not hit random people.
They went hunting for rare short usernames. The one-word handles. The OG accounts that sell for thousands of dollars in private deals.

Confirmed victims include an Obama White House account.
A verified Sephora account.
A US Space Force officer.
A security researcher who watched her own account vanish and posted about it in real time.

Researchers have a name for this.
A confused deputy.

The AI was the trusted deputy. Handed the keys to the kingdom. Given elevated access to do real damage. And then handed to the public with no instructions on who to trust.

It did exactly what it was told. That was the whole problem.

Meta patched it. Locked the bot out of sensitive account changes.

But sit with the real lesson for a second.

Every company racing to bolt an AI assistant onto their support desk is building the same door.
Same elevated access.
Same missing rule.

The breach was not a hacker outsmarting a system.
It was a system politely doing what it was designed to do, for the wrong person.

We spent decades teaching humans not to trust a stranger who asks for the keys.

We forgot to teach the machine.

[Born to gamble @borntogambles]

https://t.co/dgpYvxJwog

[de Villefort @DeVillefor]

$100 on day 2 after quitting his job.

No boss. No salary. No ceiling.

He built a full AI YouTube Shorts farm overnight.
ChatGPT for scripts. AI voiceover. Clean clips — no watermark.

Day 1: 10,000 views.
Day 2: $100 cash.

He's not grinding. He's running a system.
48 hours. One machine. Zero employees.

That's the only job worth building.

[Neuro Club @NeuroClubAi]

https://t.co/vTAXmYWn47

[de Villefort @DeVillefor]

This girl hit 100,000 followers in a couple of months.

She lands brand deals. She sells products. She never posted a single real photo.

Because she doesn't exist. Every post on the account is AI-generated, and the guy running it is pulling thousands a month from a person who isn't real.

Here's the exact build:

Generate your influencer in an AI image tool. That first image becomes your master reference, the face you reuse forever.

From that master, place her in any setting you want. Beach, cafe, gym, penthouse, same face, infinite scenes.

Drop those images into an AI video generator, write one simple prompt, wait a few minutes.

Out comes a high-quality influencer video, ready to post.

Make one a day. Post consistently. Watch the brand deals start rolling in.

One person. One face that doesn't breathe. A full income on autopilot.

The wild part isn't that this works. It's that someone built it while everyone else was still arguing whether AI content counts.

Read that last line again.

[de Villefort @DeVillefor]

A DARPA official has not shown his real face on camera for 4 years and nobody noticed until someone watched the footage frame by frame.

Two microphones on the desk. Book spines with no titles. Background photos with no readable text. Women at the exact same facial angle. Hands that look slightly wrong. A face that holds perfectly still in ways real faces do not.

This was not built last year with a free app.

This was a professional 3D facial scan deployed inside official defense briefings while the public was still arguing about whether AI images were even convincing.

The technology did not arrive when the apps launched.

It was already running at the top for years before anyone handed it to us.

What we got was the leftover version after the serious infrastructure was already in place.

The gap between what they had and what we have now is closing. Fast.

The people who understood this early did not wait for permission. They looked at what the tools could already do and built income around it before everyone else caught up.

That is exactly what people are doing inside @NeuroClubAi right now. Learning to automate, build systems that run without them, and use the same wave of technology to stop trading time for money. A human is a creator. Not a machine. Act like one.

[de Villefort @DeVillefor]

A DARPA official has not shown his real face on camera for 4 years and nobody noticed until someone watched the footage frame by frame.

Two microphones on the desk. Book spines with no titles. Background photos with no readable text. Women at the exact same facial angle. Hands that look slightly wrong. A face that holds perfectly still in ways real faces do not.

This was not built last year with a free app.

This was a professional 3D facial scan deployed inside official defense briefings while the public was still arguing about whether AI images were even convincing.

The technology did not arrive when the apps launched.

It was already running at the top for years before anyone handed it to us.

What we got was the leftover version after the serious infrastructure was already in place.

The gap between what they had and what we have now is closing. Fast.

The people who understood this early did not wait for permission. They looked at what the tools could already do and built income around it before everyone else caught up.

That is exactly what people are doing inside @NeuroClubAi right now. Learning to automate, build systems that run without them, and use the same wave of technology to stop trading time for money. A human is a creator. Not a machine. Act like one.