14:34Stockton Walbeck · Essay
Stop Selling Education. Sell These 4 Things Instead
Course revenue doesn't flatline because of AI. It flatlines because durable, one-time information was never enough on its own.
September 10thThe argument in one line.
AI video editing only avoids looking like slop when the model is given an editor's rules for sifting, storyboarding and visual layouts, plus a real review loop for revisions.
Read if. Skip if.
READ IF YOU ARE…
- A business owner paying an editor per video who wants to understand what an AI editing pipeline can realistically replace and how long it takes.
- A creator who tried Claude on raw footage, got generic AI-looking edits back, and wants to know what was missing.
- A video editor curious how editing judgment (dead space, retakes, B-roll layouts) gets encoded as rules a model can follow.
SKIP IF…
- You want a free step-by-step tutorial you can run today, since the tool shown launches November 18 behind a waiting list.
- You edit cinematic or narrative work where template-driven layouts for ads, reels and VSLs don't apply.
The full version, fast.
The creator, a video editor for over ten years, now has Claude edit his reels, ads and VSLs through a five-step pipeline: film, sift, storyboard, build and review. Sifting removes dead space and retakes using decibel gates and breath detection. Storyboarding, which he says is 80% of an edit, matches what is said to one of over a hundred layouts he reverse-engineered from tens of thousands of analyzed frames. Revisions work like Frame.io comments and return in 15 minutes to two hours. It launches November 18 as a Claude Code integration at no extra cost beyond the subscription.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 01:03
01 · Cold open and the claim
28 AI-edited videos, the cost comparison, and who the video is for.
01:03 – 01:32
02 · Step 1: Film and ingest
Phone with teleprompter and DJI mic for short form, teleprompter camera for long form, then drag everything into the editor and pick the aspect ratio.
01:32 – 02:34
03 · Step 2: The sift
Dead space and retake removal, a 6:24 raw reel cut to 76 seconds, and the rules behind it.
02:34 – 05:31
04 · Step 3: Storyboarding
Why this is 80% of the edit and how layouts for long form and short form were trained from analyzed frames.
05:31 – 05:58
05 · Step 4: The 10-step build line
Sync, transcribe, sift, storyboard, build, with a quality check between each step.
05:58 – 07:50
06 · Step 5: Review and revision
Frame.io-style comments, dropped-in screenshots, and revisions back in 15 minutes to two hours.
07:50 – 09:53
07 · Launch, cost and time
November 18 launch, Claude Code integration, 30 minutes to four hours per edit, and $20 versus $500.
09:53 – 10:01
08 · Sign-off
Join the waiting list.
Lines worth screenshotting.
- Sifting a video has two parts, dead space removal and retake removal, and doing it well takes hundreds of rules like decibel thresholds, breath detection and J-cuts and L-cuts.
- The creator says storyboarding, matching spoken words to visuals, is about 80% of what makes an edit good, whether a human or an AI does it.
- AI is not intuitively creative, so the fix is instructions: he had AI analyze tens of thousands of frames from his own and five favorite creators' videos to build over 100 templates.
- A 6 minute 24 second raw reel was cut to 76 seconds of usable footage in about two minutes.
- No edit is perfect on the first round, so a comment-and-revise loop matters more than a one-shot prompt.
- A human revision round takes at least 24 hours; the AI version returns in 15 minutes to two hours.
- He claims one video that took a human editor 5 to 10 days and about $500 took about five hours and roughly $20 of usage.
- Most AI editors fail at the sift step because their makers are not editors and never encoded how an editor decides what to cut.
- A simple edit takes about 30 minutes; a retention-maxed edit with constant B-roll and motion graphics takes two to four hours or more.
- He frames 'one take, Claude did it all' edit demos as almost certainly untrue unless you have no creative preferences.
Why AI edits need editor's rules
WHAT TO LEARN
Raw AI edits look generic because the model lacks an editor's judgment, so the work is encoding sifting, layout and revision rules, not writing a better prompt.
- Cleaning footage means more than trimming silence: decibel gates, breath recognition and retake detection decide what is dead space and what is natural pacing.
- Most of an edit's quality comes from matching each spoken idea to the right visual layout, so a library of proven layouts beats open-ended generation.
- Studying tens of thousands of frames from strong videos is a way to turn taste into instructions a model can execute reliably.
- A built-in comment and revision loop matters because creative preference is personal and first versions are never final, for humans or AI.
- Time and cost savings come from collapsing waiting, with revisions returning in minutes to hours instead of a day or more.
Terms worth knowing.
- Sift
- The cleanup pass on raw footage that removes dead space between words and breaths and cuts bad takes down to one usable file.
- Storyboarding
- Matching what is said in the video to specific visuals, such as B-roll, layouts, captions and motion graphics, before building the edit.
- A-roll
- The primary footage, usually the person speaking to camera.
- B-roll
- Supplemental footage or graphics shown over or instead of the A-roll to illustrate what is being said.
- J-cut and L-cut
- Edits where the audio of the next or previous clip overlaps the picture of the current one so the cut feels seamless.
- VSL
- Video sales letter, a long sales video built around a single offer.
- Rotoscoping
- Separating a person from the background frame by frame so text or graphics can appear behind them.
Things they pointed at.
00:33toolTeleprompter app
01:00productDJI mic
06:28toolFrame.io
08:08toolClaude Code
08:00productAI video editor waiting list (launch Nov 18)
Lines you could clip.
02:34
“About 80% of what makes a video edit, whether it's done by AI or human, comes in this one single step.”
04:30
“That is how you give AI creativity. It needs instructions.”
05:59
“No video edit is ever perfect on the first version ever.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
This video was edited entirely by AI. So were my last five Instagram reels and my last 20 ads and my last two VSLs. That is 28 videos.
At my editor's normal rates, that's $2 ,500, two and a half weeks of editing time. These came back in hours on the clod subscription I'm already paying for. I have been a video editor for over 10 years, editing YouTube videos, hundreds of ads, course content, VSLs, reels, you name it, and I have not opened Premiere Pro, CapCut, Flash...
cut pro x in two months and if you right now are paying for an editor yourself and you're tired of waiting multiple days and paying an invoice for every single video, this is for you. If you bought one of those, Claude can now edit your video from zero tools and the edit came back looking like a bunch of AI slop, this is also for you.
If you make videos for your business, this is my entire workflow start to finish for both long form and short form content in just five steps. Let me show you how this all works. Step one is I film my content.
Now for short form, I film on my phone. This is a teleprompter app that I use. I just literally pull it up right here.
I film the video and then actually, my dji mic too i film the audio from this for long form i film on my teleprompter right here and once i'm done i simply just take what i filmed the a -roll the b -roll the audio and anything else and i just take it and i drag and drop it into my ai video editor it already understands what your a -roll is it's going to grab your audio it's going to sync it all up down to the milliseconds i tell it what format i want the final video to be in whether it's 16x9 or 9x16 sometimes 3x4 depends on what the output is and then it's ready for step two as every creator knows and as all you video editors understand as well No video is ever perfect.
I am not perfect. This video already, so far, I have made a whole slew of mistakes. I've had to do multiple retakes.
There are ginormous dead spaces in between my words and my breaths. And the process of getting all that junk down to one usable file is called the sift. We are sifting a video and sifting has two parts.
Dead space removal and retake and mistake removal. For example, this Instagram reel I filmed just yesterday was six minutes and 24 seconds of raw footage. Mistakes, dead spaces, and all.
But then within just two minutes, it synced all that up. cut it all the way down to just 76 seconds of usable footage in its final state that is now ready for the next steps on this assembly line. And real quick, even something that seems as simple as just sifting a video is actually an extremely complicated science that looks at hundreds of different rules and variables like these ones here.
These are all things I intuitively knew as a video editor, but never trained an AI model on. You have to look at all the DB levels to determine what is a dead space and what isn't. You have to know what a breath is and what that sounds like, knowing what a retake is, using Jcuts, L -cuts to make sure that they all seamlessly sift together.
This right here is why so many other AI video editors out there completely bought this step of sifting your video. The people who made these tools are not video editors by heart, and so they don't understand how to engineer an actual video editor to begin with.
But once everything is all sifted, it's now ready for the most important step of the entire video editing process, and that is storyboarding. This is where all the magic happens. About 80 % of what makes a video edit, whether it's done by AI or human, comes in this one single step right here.
This is the creative process of trying to take what is being said verbally in a video and matching that with visuals. If there's one thing I've learned about AI, it is not a very creative thing, right?
AI just looks at ones and zeros. It looks at a big set of code. It has to have instructions.
It doesn't quite yet have that intuitive, creative nature that humans have. That is, unless you train it, right? Because how is AI supposed to know how to lay out the B -roll on the screen?
What to do with your captions, the motion graphics, the title animations, the sound designing? It doesn't. It has to be trained.
For long form content like this YouTube video or for a course video or for VSLs, sometimes I want one big word behind me with my body rotoscoped behind it. Looks nice and clean. Sometimes I want to compare two things with a list on my right and a list on my left with me in the middle.
Sometimes I want visuals to match bullet points, like when I make lists, right? Step one, step two, step three, and so on. Sometimes I don't want you to look at me at all and just show the full screen to be visuals only.
Let's show four main points or maybe multiple examples or just a full custom -made motion graphic. Sometimes I want... this A -roll to get blurred out and then have words overlaid on top of that.
And then for short form, it needs a whole separate set of layouts and rules and instructions and science behind it. Sometimes I want a split screen layout with me on the bottom and then the thing on top captions in the middle. Sometimes I want to have three things pop up above my head to show the visuals I'm talking about.
Sometimes I want myself as a green screen on the bottom part while I highlight and talk about something above me. Sometimes I want to go full screen on the B -roll, just showing some animations or some graphics. You get the idea.
But the point is, this has all been scientifically reversed engineered. I've taken every one of my YouTube videos, Instagram videos, ads.
I've taken my top five creators who I like. And I had AI analyze literally every single frame, tens of thousands of them to understand what actually is happening on the screen. How does it animate into the screen to create over a hundred of the most used templates to take what is being said verbally and match it with the best possible visual on the screen.
That is how you give AI creativity. It needs instructions. What came back after days and hundreds of thousands of frames being analyzed, later, is a big, ginormous library of every visual move you can see in the best videos online.
Every single one of these has a big set of code. So AI knows exactly how to make it come to life in an actual video. And this library will only continue to grow every single month as new trends and new visuals come about.
And then once everything's done being storyboarded, all of the verbal things that you're saying are being matched with their visuals. Then it comes to the building phase. I designed my AI video editor to work through a 10 step assembly line.
It's kind of like McDonald's, but for video edits, right? The same 10 step process. I would use myself as a video editor and that I trained hundreds of my video editors in my video editing agency to use.
You can see this video right here working through each of those 10 steps. The syncing process, the transcribing of the video, the sifting, timestamping every single millisecond of the video, the storyboarding process, the building, and each step after getting done has to get quality checked before it moves to the next. Which leads us to the fifth and final step of having a video get fully edited by AI.
It is the review and revision process. A quick news flash for you. No No video edit is ever perfect on the first version ever.
At this point, I have gone through thousands of edits that have been done by my human editors. And guess what? I don't think there's even a single time where it was perfect on the first round.
And that's kind of obvious why. We are creative creatures. The person editing, whether it's AI or human, just has different preferences and different creative decisions they made.
And you will not always agree with those decisions. And so I designed a built -in revision and review process modeled after what you'll see inside of like frame. So just like my human editors would send me back the final edit, I get to go through, leave all of my comments, and I just get to pick it apart.
I leave comments about different decisions made on the B -roll. Maybe I didn't use certain colors or the captions weren't right. Maybe I've got my own creative idea of what I want to go here in the video.
And so I'll leave my thoughts, then I'll comment on it. I can drag and drop screenshots or different files directly into here, and AI can analyze all of those. The difference, however, is that a revision for an edit like this one, would take at least 24 hours to get back to me on my desk.
That's the fast case. Because again, we as humans, we take breaks, we have shifts, we go to bed, we wake up, and then we'll work on the edit tomorrow, right? AI knows no such limits.
And so it just goes ahead, it applies revision, and anywhere from 15 minutes to two hours later, depending on the complexity of your feedback, it's going to give you back a beautiful revised version. But the beautiful part about why this AI tool is so easy to give revisions to is because it's already trained on all that creativity, all the templates, all those...
decisions already made, you just have to guide it. The AI does the rest and it makes all those visuals come to life. So all those videos you might be seeing where people are saying, this edit was done in one take, one shot, Claude did it all.
It's amazing. First of all, almost certainly not true. Second of all, that's really only the case if you have no creative preference whatsoever, which that might be you.
Maybe you're just happy with what it made on round one. But if you're more of an artist, kind of like I am, then you've got a lot of creative preferences and you want to fine tune it and tweak it to your... taste and style.
So this whole five -step process to get a video from my phone or my camera and fully edited, start to finish, how long does it take? How much does it cost? And how can you get access to it?
Everything I just built here today is launching on November 18th. If you know anything about me, I am very careful to release products with my name and reputation behind them. I want to make sure that we get beta users, that it runs through multiple versions until I think it's darn near perfect before I let you get your hands on it.
So you can join the waiting list for the second I make it go live and drop it on November 18th. This is not an A. AI slop vibe coded piece of software.
This took multiple months to get here. And I am ecstatic to get into your hands because it is saving me hundreds of hours every single week and thousands of dollars every single month and paying for video editors. How it works, it's a custom integration that plugs into your cloud code.
I'll probably make it available for Codex and other AI softwares in the future as well. It loads up the entire user interface to actually use the video editing tool. And so in terms of costs, it costs you nothing more than the cloud code subscription you're currently using.
In terms of time right now for a very simple edit, it takes about 30 minutes to get it done. If, however, you are somebody who likes a more punchy, very almost like retention maxed video with motion graphics and B -roll popping up every single second, that's a lot more time to edit. That might take two to three, sometimes four hours plus.
Again, depending entirely on the length of the content, if it's a short form video and other variables. But for me, the magic of this all is comparing it to what I was doing before, right? This one video right here would take me on average five to 10 days to get back from my video edit.
At the quality that I like with all the revisions included, that's about how long it would take me. And it would cost me on average about $500 to get this. This video here took about five hours to edit and probably about $20 worth of usage.
That... is astronomical. And that is precisely why I have been spending the last three months of my life condensing 10 years of video editing knowledge built specifically for creators to build course content, long form YouTube content, Instagram reels, TikTok videos, YouTube shorts, VSLs, you name it, whatever your editing need is, my goal is to accomplish both of them.
And you can get early access by joining the waiting list below. And that folks is how I'm using AI to edit all of my content in 2026 and how in 2027, it's going to be editing 100 % of the content I do. That's it for this video.
I'm really excited to release this. Jump on the waiting list and I'll see you in the next one.
The Hook
The bait, then the rug-pull.
The pitch is blunt: 28 videos, $2,500 of editing and two and a half weeks of waiting, replaced by hours on a Claude subscription. The rest of the video explains the five steps that supposedly keep it from looking like AI slop.
Named ideas worth stealing.
01:03list
The five-step AI editing workflow
- Film
- Sift
- Storyboard
- Build
- Review and revise
A pipeline from raw phone or camera footage to a finished edit, with a quality check between each stage.
01:32model
The sift
- Dead space removal
- Retake and mistake removal
Uses decibel levels, breath detection, retake detection and J-cuts and L-cuts to reduce raw footage to one usable file.
How they asked for the click.
VERBAL ASK
08:00product
“Join the waiting list below for the November 18 launch.”
Held until after the five steps and the cost comparison, then repeated at the sign-off.
Visual structure at a glance.
