Your YouTube AI Assistant To Create More Videos And Save 10 Hours A Week
Joined June 2024
- Tweets 701
- Following 10
- Followers 4,580
- Likes 250
272 Photos and videos
Pinned Tweet
Mar 23
Introducing Thumbnail Studio.
Turn any Subscribr script into a scroll-stopping thumbnail in under 60 seconds.
Set it up in 60 seconds and it gets to work. Brainstorm concepts, clone competitor styles, preview next to rivals, auto-improve until it's right.
Available today at subscribr.ai
13
6
89
12,077
Life if you avoid these niches...
3 YouTube niches I would NEVER start a faceless channel in right now.
1/ AI stories.
This is the niche getting wiped first in 2026.
Reddit story narrations. AI sleep stories.
Templated fiction reads with the same 10 stock images on loop. Same AI voice. Same plot structures.
The audience is either unconscious or scrolling, which means advertiser ROI is zero.
YouTube hunts this format specifically because the platform cannot make money on it.
Mass demonetizations are happening every week. The math is broken at the platform level.
2/ News and politics.
The numbers look big. The reality is brutal.
RPMs sit at $0.50 to $2 because advertisers refuse to run next to political content.
The audience is highly engaged but worth almost nothing in CPM terms.
Election cycles bring traffic spikes that crash to zero within 30 days.
Demonetization happens for any controversial take. One misread headline and the channel is buried.
Massive engagement, almost no revenue, constant flagging risk.
3/ World Cup and tournament-specific soccer content.
Seasonal niches are operator graveyards. World Cup, Euros, Copa America. The traffic spikes for 4 weeks every 2 to 4 years. Then it dies completely.
You spend 18 months building an audience for content that has a 4 week monetization window.
The rest of the year your channel pulls 500 views per upload.
The pattern across all 3 dead niches.
Low RPM. Advertiser-unfriendly content.
Seasonal or templated formats the classifier hunts.
High copyright risk. Spiky traffic that collapses.
Every operator running these niches in 2026 is paying the platform to host their work for free.
Skip them.
Pick an evergreen niche with older Tier 1 audience, high RPM, and unlimited topic supply.
The dead niches die slow. The smart operators leave early.
1
179
The smartest AI avatar move on YouTube right now.
Stop hiding the AI. Start announcing it.
Julia McCoy runs a 200K subscriber AI news channel with a 100% cloned avatar of herself.
Inside every video, the avatar says it directly. "This is an AI clone of Julia McCoy, while the real Julia is helping change lives with her AI company."
The audience does not care.
The algorithm does not flag it. The channel posts daily, pulls 10K to 50K views per upload, and lands brand sponsorships consistently.
The lesson is brutal once you see it.
Trying to hide the AI is what gets channels demonetized.
Being upfront about it builds trust and works inside the platform's own rules.
If you are running an AI avatar channel in 2026, write the disclosure into the script itself.
Put it in the description. Add it to the channel about page. Make the AI a feature, not a secret.
The platform rewards transparency.
The classifier punishes deception. Pick the right side of that line and your channel survives anything.
1
11
575
How to read your YouTube analytics in 47 seconds.
YouTube Studio gives you 50 metrics.
47 of them are noise.
Only 3 dashboards actually decide what your next video should be.
1/ The CTR vs APV scatter plot.
Open Studio. Click Analytics. Switch to Advanced view.
Pull up the chart that maps Click Through Rate against Average Percentage Viewed across your last 30 uploads.
Top right corner. High CTR and high retention.
These are your winning formats.
Make 10 more of them. Reverse engineer the title structure, the thumbnail style, and the topic angle. Copy your own wins.
Bottom left corner. Low CTR and low retention. Dead formats. Cut them and stop revisiting the topic.
Top left. High retention but low CTR. The video is great. The packaging is broken. Rewrite the title, redesign the thumbnail, and republish the format with a sharper hook.
Bottom right. High CTR but low retention.
The packaging is bait and the script does not deliver. Either tighten the writing or kill that title style entirely.
90% of the decisions you need to make live inside this one chart. Most operators have never even opened it.
2/ The Returning Viewers chart.
The single most underrated metric on the platform.
It tells you what percentage of your viewers came back to watch a different video on your channel in the last 28 days.
If the number is climbing, the channel is healthy.
The audience is becoming a real audience.
If the number is flat or sliding, your videos are not landing.
New viewers are showing up once and disappearing.
You are running a one-shot lottery instead of building a media business.
Most operators obsess over total views. The pros obsess over returning viewers.
3/ Average Views Per Viewer.
How many videos the average viewer watches in a single session.
This is the metric YouTube's algorithm uses to decide how hard to push your channel.
A viewer who watches 1 video and leaves makes you a dead end.
A viewer who watches 4 makes you a session driver.
YouTube pushes session drivers because they keep users on the platform longer.
That is the entire reason the recommendation engine exists.
How to lift it. End screen CTAs that point to the next logical video.
Playlists that auto-progress through your best content.
Series uploads with open loops that pull viewers from one video into the next.
That is the entire weekly analytics review.
2
9
295
3 YouTube niches I would NEVER start a faceless channel in right now.
1/ AI stories.
This is the niche getting wiped first in 2026.
Reddit story narrations. AI sleep stories.
Templated fiction reads with the same 10 stock images on loop. Same AI voice. Same plot structures.
The audience is either unconscious or scrolling, which means advertiser ROI is zero.
YouTube hunts this format specifically because the platform cannot make money on it.
Mass demonetizations are happening every week. The math is broken at the platform level.
2/ News and politics.
The numbers look big. The reality is brutal.
RPMs sit at $0.50 to $2 because advertisers refuse to run next to political content.
The audience is highly engaged but worth almost nothing in CPM terms.
Election cycles bring traffic spikes that crash to zero within 30 days.
Demonetization happens for any controversial take. One misread headline and the channel is buried.
Massive engagement, almost no revenue, constant flagging risk.
3/ World Cup and tournament-specific soccer content.
Seasonal niches are operator graveyards. World Cup, Euros, Copa America. The traffic spikes for 4 weeks every 2 to 4 years. Then it dies completely.
You spend 18 months building an audience for content that has a 4 week monetization window.
The rest of the year your channel pulls 500 views per upload.
The pattern across all 3 dead niches.
Low RPM. Advertiser-unfriendly content.
Seasonal or templated formats the classifier hunts.
High copyright risk. Spiky traffic that collapses.
Every operator running these niches in 2026 is paying the platform to host their work for free.
Skip them.
Pick an evergreen niche with older Tier 1 audience, high RPM, and unlimited topic supply.
The dead niches die slow. The smart operators leave early.
2
529
3 YouTube niches I would start a brand new faceless channel in right now.
1/ AI doctor and health for the 55 crowd.
Highest RPM on the platform. Older audience with disposable income that advertisers chase with premium ad spend.
Evergreen topics that never die.
Easy to layer in a $27 ebook or longevity guide on top of AdSense.
Avatar channels in this space monetize fast and almost never get flagged because the educational value is obvious to the algorithm.
2/ AI financial advisor and retirement planner.
Same older Tier 1 audience. $4 to $7 RPMs consistently.
The audience watches 20 minute videos start to finish and pays for the consultations afterward.
The US is saturated. The UK, Canada, and Australia are wide open right now.
3/ 4K travel documentaries. Country breakdowns.
$5.54 RPM. 30 to 45 minute videos that pack in more ad slots.
Older audience that watches every second.
Real footage based which means almost zero risk of inauthentic content flags.
The common thread across all 3.
55 Tier 1 audience. High RPM. Evergreen topics. Real value per video. Low demonetization risk.
If you are picking a niche this month, pick one of these. Commit for 12 months.
The math works. The data works. The audience is already there.
Stop picking niches because they sound cool. Pick the ones the algorithm pays.
3
2
15
646
Jun 13
You in 2 years if you don't watch this...
Jun 9
How to Start an AI Avatar Channel From Scratch (Full Step by Step Guide)
4
274
Jun 13
POV: You have an AI avatar channel in 2026
Jun 9
AI avatars just did to YouTube what the iPhone did to phones.
When the iPhone launched in 2007, almost nobody understood what was happening.
The early adopters built app empires worth billions.
The late adopters spent the next decade explaining why they were not on it yet.
AI avatars are the same shift, happening in 2026, and the window is open right now.
Here is what changed.
Before AI avatars, building a YouTube presence required 4 things at once.
The time to film weekly. A camera setup that did not embarrass you.
An editor who could turn raw footage into something watchable.
The confidence to actually sit in front of a lens and talk for 10 minutes without freezing.
Almost no professional, business owner, or expert had all 4 at the same time.
So they did not start. The smartest doctors, lawyers, dentists, and consultants in the world stayed invisible online for an entire decade because the barriers were too high to push through.
The technology just removed every single one of those barriers in 18 months.
You record 5 minutes of yourself once.
The system trains an avatar that looks indistinguishable from a real recording of you.
From that moment forward, you feed it scripts and it ships videos with your face and voice.
No camera setup. No editing crew. No weekly filming sessions. No nerves.
The avatar shows up to work even when you do not.
The iPhone parallel matters because the same dynamic is playing out in real time.
In 2007, the smart move was to build an app the moment the App Store opened. Instagram, Uber, Tinder, WhatsApp, Snapchat. All built in the early years when most people were still asking whether smartphones were a real thing.
In 2026, the smart move is to launch an AI avatar channel the moment the technology became indistinguishable from real recording.
The professionals who do it now will own their niche for the next 10 years. The ones who wait until 2028 will discover the category is already locked up by the early movers.
Doctors with avatar channels are about to dominate health YouTube. Lawyers with avatar channels are about to own consumer law search.
Dentists with avatar channels are about to corner every cosmetic procedure query.
Financial advisors with avatar channels are about to take over retirement planning content for the next 20 years.
None of this requires capital. None of this requires technical skill. None of this requires you to suddenly become comfortable on camera.
It requires the decision to act before the window closes.
The iPhone era made the first wave of app founders rich. The avatar era is going to make the first wave of professional creators rich in the exact same way.
464
Jun 12
I launched a brand new AI avatar YouTube channel in 45 minutes on a live call.
Banner. Profile. Avatar. Voice clone. Niche.
Channel description. Ideation. Script. Thumbnail.
Tags. Description. Scheduled upload.
All inside one platform. One operator. One sitting.
10 years ago this would have required a full production team, a videographer, a script writer, a designer, and 2 weeks of work to ship the first video.
In 2026 it takes one cup of coffee.The bottleneck for starting a YouTube channel is officially gone.
The only thing left between you and your first upload is the decision to start.
Stop "researching." Open Subscribr. Ship the first video this weekend.
1
12
713
Jun 12
The math nobody runs on AI script tools.
Connecting your own Claude API key to Subscribr costs roughly $1 to $2 per script using Sonnet.
If you write 8 to 16 scripts a month, that is $8 to $32 in API costs total.
Compare that to writing the same scripts manually.
4 to 8 hours per script. 32 to 128 hours of writing per month.
At any real hourly rate, that is thousands of dollars of opportunity cost burned every month.
Or compare it to hiring a scriptwriter. $200 to $500 per script. $1,600 to $8,000 a month for the same output.
The Script Agent with your own Claude API key replaces both. For pocket change.
Most operators are still writing manually because they think AI scripts cost more. The actual math is brutal.
AI scripts are 99% cheaper than every other option. The bottleneck stopped being the writing.
It stopped being the cost. The only thing left in the way is the decision to use the tool.
9
574
Jun 12
I just sent a voice note into Subscribr and got a 4,000 word script back.
No outline. No research. No structure work. No editing.
The agent took my voice note, did the research, filled in the gaps, and built the full script around what I actually said.
This is the workflow most operators have not figured out yet.
You do not need to type a perfect prompt. You do not need to write a detailed brief. You just talk into your phone for 5 minutes about the topic.
The agent picks up your perspective, your opinions, your structure, and your voice.
Then it fills in the research, the supporting data, and the script polish.
A 4,000 word script from a 5 minute voice note.
Talk on the way to work. Get a finished script before you arrive.The era of "I do not have time to write" is officially dead.
8
460
Jun 11
The 9 rules every YouTube thumbnail text follows in 2026.
1/ Start with no text. Only add if it improves the thumbnail.
2/ Max 5 words. No exceptions.
3/ Never restate the title.
4/ Highlight the outcome, the problem, the twist, or the proof. Not the topic.
5/ Communicate 1 core idea. If you cannot say it in 5 words, the idea is fuzzy.
6/ Readable on mobile. Pass the 10 foot squint test.
7/ Design for psychology, not decoration. Bold sans serif. High contrast. Max 3 colors.
8/ Place text away from the face. Text left, face right. Or the reverse.
9/ Order of importance. Face first. Text second. Everything else last.
Less text. More click.
2
182
Jun 11
The 8 question worksheet I run before designing any YouTube thumbnail.
1/ Is this a new upload or an A/B test?
2/ Who is this video for? Define the exact audience.
3/ What are the pain points of that audience for this specific title?
4/ What are the desires of that audience for this specific title?
5/ What is the angle of the video and thumbnail?
6/ What are the title options? Build the thumbnail around the strongest one.
7/ What 2 to 3 principles will you focus on inside the thumbnail?
8/ What 3 main elements will you include, and how will you present them visually?
1
205
Jun 11
The fastest way to find a winning YouTube thumbnail in 2026.
Start with 6 concepts. End with 1 winner.
Step 1. Generate 6 thumbnails. Different designs. Different emotions. Different compositions. Quantity first, polish second.
Step 2. Pick the 3 strongest. Drop the bottom half immediately.
Step 3. A/B/C test the 3 concepts on YouTube. Watch CTR for 48 hours. The data picks the winner, not your taste.
Step 4. Lock the winning concept. Now test the text overlay. 3 variations. Same image. Different words. A/B/C test again.
Step 5. Lock the winning text. Now test 3 titles paired with the thumbnail. A/B/C test one more time.
The output is a thumbnail and title combination that has been validated 3 times against real viewer behavior.
Most creators ship the first thumbnail they like. The pros ship the one that won 3 separate tests.
The math is brutal in favor of the operators who run the funnel.
1
162
Jun 11
Every great YouTube thumbnail has exactly 3 elements.
1. The face. Emotion the viewer can read in 0.5 seconds.
2. The visual hook. A map. A chart. A product. A diagram. The thing that gives the topic context.
3. The text. 3 words or less. Curiosity-driven. Specific.
That is it.
3
16
655
Jun 11
The 80/20 of avoiding YouTube demonetization in 2026.
Stop committing the 8 patterns the inauthentic content classifier hunts.
Robotic voice. Repeated title formats. Identical thumbnails.
Same paragraph rhythm. No real research. Default stock voiceovers. Visual sameness.
Content optimized for ads instead of value.
Commit 4 or more of these and you enter the suppression zone.
Most operators commit 5 without realizing.
The fix is variation. Custom voice. Rotated thumbnail layouts. Varied title formats.
Real research per video. Mixed visuals. Educational depth.
This is what separates the faceless channels surviving in 2026 from the ones getting wiped every week.
Vary every layer.
Disclose AI usage. Add real value. Survive any wave.
The classifier does not hate AI.
It hates the operators who refused to put effort into it.
1
1
3
208
Jun 11
60% of internet content is fully AI generated right now.
By the end of 2026 that number hits 90%.
YouTube is not the enemy of AI.
The platform that made Gemini, Veo, and Notebook LM cannot afford to be.
Every talking head video on the platform already uses AI somewhere in the stack. Scripts. Research.
Thumbnails. Edits. The platform knows it and has no problem with it.
What YouTube actually hunts is bad AI usage. Mass-produced templates. Repetitive voices. Stock visuals on loop. Zero value to the viewer.
The fingerprint of slop is what gets demonetized. Not the AI itself.
If you use AI to research faster, write sharper, ship more, and add more value per video, the algorithm rewards you.
If you use AI to flood the platform with templated trash, the algorithm finishes you.
The technology is not the problem. The intent is.
Build with AI.
Add value with humans. Both can coexist on the same channel without ever getting flagged.
2
1
13
822
Jun 11
The smartest YouTube move in 2026 is starting 3 channels at once.
The math nobody runs.
1 channel. 1 video a week. 4 videos in month 1. 1 trust score test. 1 niche test. If anything is broken, you wasted 30 days.
3 channels. Same 1 video a week per channel. 12 videos in month 1. 3 trust score tests. 3 niche tests.
Same time investment because the AI stack runs all of them with the same scripts, thumbnails, and editing workflow.
By day 30 the signal is clear. Kill the channels that flatlined. Double down on the one that popped.
The 1 channel approach makes you a hopeful creator. The 3 channel approach makes you an operator with a real portfolio strategy.
Most operators stop at 1 because they think the workload scales linearly. It does not.
Stop running 1 channel. Run 3.
1
19
889
Jun 11
Pick the audience before you pick the niche.
The single biggest mistake YouTube operators make in 2026.
They pick a niche they personally find interesting, then wonder why the channel never makes money.
The audience decides the RPM. The niche is downstream.
The highest-paying audience on YouTube is 55 in Tier 1 countries. US, Canada, UK, Australia.
Why this audience prints money.
Editing demands are low. Older viewers do not need pattern interrupts every 8 seconds.
A clean talking head with simple B-roll outperforms a hyper-edited video aimed at 20 year olds.
Retention is naturally higher. Older audiences watch start to finish if the hook delivers. Younger audiences swipe the second their attention wavers.
Returning viewership compounds. Older audiences stay loyal to the same channel for years. Younger audiences forget about you in 30 days.
RPMs are 3 to 5x higher because advertisers actually want to reach this demographic.
Pick the 55 Tier 1 audience first. The niche follows naturally. Real estate, financial advisor, retirement planner, AI doctor, AI lawyer, political news, religious content. All print.
Build for the audience that pays. Not the audience you find interesting.
2
23
875
Jun 11
3 ways to find a working YouTube niche in 5 minutes.
Copy from Instagram or TikTok.
AI avatars are blowing up on short form platforms with no YouTube version yet.
Find one with 500K followers. Replicate it on YouTube. If it worked there, it works here.
Steal a working niche and change the country.
A US channel pulled 1.2M views on senior bills.
Someone copied it for Canada within 24 hours.
125K views in 5 days. Same script, different country. Translate any US outlier to UK, Canada, or Australia.
Add a face to a working faceless niche.
Every major faceless category eventually gets an avatar version. Health became AI doctors.
Finance became AI financial advisors.
Political news became AI commentators.
Find a working faceless niche. Add an avatar. Leapfrog 80% of the operators in that space.
3 methods. 5 minutes. Pick one. Commit for 12 months.
The hard part is not finding ideas. The hard part is execution.
1
7
24
1,264