YT YouTube2Text

YouTube2Text → Transcripts → His EXACT System For Writing GREAT YouTube Scripts With AI

His EXACT System For Writing GREAT YouTube Scripts With AI — Transcript

by Jay Clouse · 10,620 words · 1,441 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Where is it worth putting my energy into
0:01learning to use AI well? And where
0:04should I deliberately take a stand and
0:06be like, "No, I'm using my brain here
0:07because it's actually better to do so."
0:09Meet George Blackman. He started as a
0:12full-time script writer for Ali Abdal.
0:14He's written scripts for creators from
0:1630,000 to over 10 million subscribers,
0:18and he now teaches creators how to write
0:20their own scripts that keep people
0:22watching and turn them into paying
0:24customers. In this episode, you'll learn
0:26his exact system for writing scripts
0:27with AI, starting with the framework
0:30behind every hook he writes, for
0:32educational hooks. I always think target
0:34transformation stakes. He's also
0:36brutally honest about the ways AI lies
0:38to you in the process. You sort of learn
0:40all these little ways in which it may
0:42appear to be doing something well and
0:44then the output that comes out, it
0:45actually reveals, oh, hang on, it wasn't
0:47doing these like 10 things that it
0:48claimed very confidently that it was.
0:49and stick around to the end where George
0:51shows me the tool he built for writing
0:53killer scripts using AI that saves you
0:55time without sacrificing your unique
0:57voice. Thank you to one of 10 for
0:59sponsoring this video. As just a a frame
1:02for the overall conversation, some may
1:05think that okay, because we have AI,
1:09is that doing all of the work of script
1:11writing now? What say you? I think this
1:13is a place that everyone is now in with
1:16pretty much any industry whether it's uh
1:18to do with writing specifically whether
1:20it's coding whether uh it's like
1:22anything that could feasely feasibly be
1:24affected by AI there is this sense of
1:26uncertainty and like where should I put
1:29my energy right now? Should I be putting
1:30my energy into uh like getting really
1:33good at ideiation because the script can
1:35just be written for me or is ideation
1:37already off the cards cuz AI can just do
1:39that for me? um do I even need to bother
1:41learning to code anymore because AI can
1:43just do that for me? And I think script
1:45writing is such a broad church. There
1:47are so many aspects to it that even with
1:50that it's like okay sure AI could write
1:52my whole script but should I still be
1:54the one looking at my hook because I
1:55know the hook's super important or are
1:57these like key moments where I call out
1:59to a product that I'm trying to to sell
2:00or something in the script. Are they so
2:02important that I should still be using
2:03my brain or is AI better than me? And so
2:05more broadly like we're gonna we're
2:06going to talk about the sort of process
2:08that I use these days to write a script
2:10using AI. But I think the broader meta
2:13skill and the thing to try and figure
2:15out and that I'm trying to figure out
2:16and I think we're all trying to figure
2:16out is like where are the parts where
2:18it's the most useful? Where is it worth
2:20putting my energy into learning to use
2:22AI well and where should I deliberately
2:25take a stand and be like no I'm using my
2:27brain here not through some like moral
2:30crusade but because it's actually better
2:31to do so. Can you walk us through, let's
2:34imagine you are making a video for the
2:36George Blackman channel. How are you
2:39going about script writing today? The
2:41first step and this is where the the use
2:44case for AI is is much much stronger. I
2:47think even just to re refer back to your
2:49previous question, it's like where are
2:50people needed? It's almost easier to to
2:52list the few places where AI is
2:54definitely a no-brainer. And the first
2:56step of my process these days is very
2:58much where it is a no-brainer, which is
3:00just explaining my thoughts about the
3:02script that I want to write verbally
3:04using um either like a voice note thing
3:06or or just directly into Claude. Just
3:08that ability to spiel off in natural
3:10language what I think about a topic,
3:12even if it just allows me the chance to
3:16uh clarify my thoughts on something that
3:17I've never said out loud. Kind of the
3:19same idea as as journaling or something.
3:20It allows you to to just get more out
3:22than if you're just vaguely thinking
3:23about something in your head. And is
3:25this assuming that you've already
3:26decided on the video idea if not
3:29package? Yes, this is assuming you've
3:30already you've already thought of the
3:32packaging or I guess yes, ideally the
3:34packaging. Not everyone does it. There's
3:35no shame if you don't come up with your
3:36packaging first, although obviously try
3:37to, but yes, the idea is already decided
3:40at this point. I guess we could talk
3:41about ideas as well, but that's less
3:42where I'm involved. Um, this is I know
3:45what I'm going to talk about. Maybe I
3:47don't know exactly how I want to frame
3:48it. Um, an example from my channel
3:51specifically was like I wanted to do a
3:53video about my sort of journey in my
3:57business, but I knew I needed to tie it
3:59back to YouTube script writing in some
4:01way. And so I spoke out loud to an AI
4:04about this and I was like, here are the
4:05the sort of points in my business when I
4:07was a script writer, then I was like a
4:08consultant, now I run this community and
4:09blah blah blah blah blah, but what's the
4:12sort of angle that makes this
4:13interesting for my audience who cares
4:15maybe a bit about business but more
4:16about script writing? And what it
4:19allowed me to do was get those thoughts
4:20out and then have some time going back
4:22and forth on the angle that we we should
4:25approach the subject. And ultimately we
4:27kind of landed on this approach of at
4:29every stage of my business, how have my
4:32thoughts about script writing changed
4:33and like what's the key lesson about
4:34script writing I learned at each revenue
4:36level or or whatever which kind of um
4:38yeah tied it back to my audience's
4:40interests.
4:41>> Yeah. Let's let's imagine that somebody
4:42watching this doesn't have any type of
4:44complex projects set up. Maybe they've
4:48never used Claude or Chat GPT to do
4:52anything like this. And so what should
4:55be the starting point? Is it just like
4:57dumping this in or do you think they
4:59should take some time and say like,
5:00"Okay, here's my audience. Here's this.
5:02Here's that." Like what context does
5:05this thing need before you start talking
5:07about a particular video, if it needs
5:09context at all? So, first thing I
5:11currently use Claude. I used to use
5:13ChatgBT. I switched to Claude because I
5:15wasn't loving ChatgBT's outputs. Since
5:17I've switched to Claude, I think uh it
5:20seems like Chat GBT is then has had a
5:22kind of resurgence at the time of
5:23recording. I think it's on 5.5 now, is
5:25it? Or is it something else? Anyway,
5:27sidetrack. I'll use Claude. Either one
5:30is fine is what we're hearing. Isn't it?
5:31>> Either one is fine. Yeah,
5:32>> you'll probably have workable good help
5:35using either one of these things.
5:36>> Yeah, exactly that. Um, but in terms of
5:39giving the AI enough context, I think
5:41there's really like an easy route and a
5:42hard route. So the easy mode of this I
5:44would say is you need a style guide and
5:48you need an audience avatar and arguably
5:50then some project instructions like
5:52telling the project what it is for.
5:54That's really the the simplest thing.
5:56How do you make those? Again, it's very
5:58simple. This is that meta skill again
5:59where it's like you you explain to it
6:01this is what this project is. I need a
6:04style guide. I need an audience avatar.
6:06What do you need from me? So you can
6:08create those for me if that makes sense.
6:11How many scripts of mine or like
6:12examples of writing or if you don't have
6:14any examples of writing you would say
6:15like can I just speak to you for 5
6:18minutes about a random topic and you get
6:19my voice from that. But th those would
6:21be the two things that um you can just
6:23ask your LLM for help making. So that's
6:25the easy mode. I think the advanced mode
6:28shall we call it and this is how I I
6:30start my projects these days is same
6:32thing starting a new project with um
6:35whisper flow or just you know anything
6:36that you can you can leave a voice note
6:38to and just explaining this is the the
6:42task that I have before me. So I did
6:43this when I was upgrading my uh my
6:46course. I was like I need to upgrade my
6:48course. My course is 36,000 words long.
6:50Since releasing it I have uploaded 2
6:53years worth of newsletters. I want to
6:54integrate those newsletters into my
6:56course. I have XY Z other concerns.
6:5936,000 words is quite long. Will you be
7:01able to handle that? Just giving it all
7:03the context that I thought it might
7:04need. And then basically saying to it
7:06like what is my first step here? What do
7:07you need from me? And I think people
7:10underestimate this approach a lot
7:12because it kind of feels like ah I guess
7:13I need to know how to use this thing
7:16before I start using it. But the joy of
7:18it is and why it feels so kind of
7:20weirdly addictive is you can just say
7:21like I don't know what I'm doing. like
7:22what do I how do I how do I proceed?
7:24This is legitimately like what everybody
7:27is selling in the world of teaching you
7:30how to claude. By the way, it's like I
7:31did this thing. Pay me $99 and I'll show
7:34you how I did it. And it inevitably
7:35comes back to and then you ask it how to
7:38do it and it will tell you like that's
7:40that's the entire secret of working with
7:42AI is like just start with I'm trying to
7:45do this thing. What do you need from me?
7:47Literally, if you counted how many times
7:48I asked my my open claw, "What do you
7:51need from me?" It would be astronomical.
7:54Yes. Yeah. 100%. And and like recently
7:57building this app, I was suddenly aware
7:58like, "Hang on, there are probably some
7:59legal ramifications of selling a
8:02subscription thing that relates to AI."
8:05And so I just said to it like, "What
8:06what are the legal ramifications?" And
8:08then this enormous block of text came
8:10out and it was like, "We have these 25
8:12tasks to do. You need to contact this
8:13lawyer. You need to blah blah blah
8:15blah." And suddenly it's like, oh yeah,
8:16okay. I mean, I didn't know how to do
8:18that, but but now I do. Either mode or
8:20hard mode, you get up, you set up the
8:22context for your voice guide, your
8:24audience avatar, and now you start
8:26saying, "Okay, now that I've zeroed in
8:28on a video idea, I'm going to just brain
8:31dump voice dictation or type whatever
8:34pleases you, what I'm thinking about
8:36this video." How do you give it the
8:38instruction for the next step? The next
8:41kind of thing that I like to have is
8:43something that trains it on what a
8:45YouTube script looks like in relation to
8:47other types of writing. You can totally
8:49build these yourself. You could even
8:51just take a bunch of transcripts from
8:53your own channel, from channels in your
8:54niche that are doing well. Not to say
8:57steal how they do stuff, but like
8:58structurally, what what what do you
9:00detect about the structure of a YouTube
9:01script? And how can we use that to train
9:04you on what a script should look like? I
9:06have a free thing that has my version of
9:08what YouTube structure looks like. It's
9:10in the description right now. That's
9:11kind of the next thing. And because I
9:13think then it saves you the hassle of
9:14having to be like to to say here is my
9:17brain dump of thoughts now structure it
9:19like a YouTuber would. It kind of has
9:22this pre-loaded context of what a
9:23YouTube script would look like. So that
9:25that's really the next step is once
9:27you've brain dumped your stuff, you want
9:30to either through the pre-loaded
9:32training or saying to it outright, let's
9:34first of all figure out what the
9:35overarching structure of this would look
9:37like. and it's going to come back to you
9:39with a suggestion and you you see if it
9:41it fits and go from there. For people
9:43who are stubborn and haven't downloaded
9:44this free thing yet, what are like the
9:46highlevel bits of this in there that
9:50people should know is like part of this
9:52structure. Three main blocks, shall we
9:56call them, um that I would think of for
9:57a YouTube script. The hook is is the
9:59obvious first block. Uh it's quite
10:02distinct how you would go about writing
10:03a hook. There are different frameworks
10:05that different people have for this. For
10:06me, for educational hooks, I always
10:09think target transformation stakes. So
10:11that is who is my target, who is my
10:12audience. What is the transformation
10:14this video offers and what is at stake
10:16if they don't achieve that
10:17transformation? The second main block is
10:19is a segment. So these are your
10:21subtopics within the video. For every
10:23segment of my video, I think of it in
10:25terms of setup, tension, and payoff. So
10:27at the start of each segment, we set up
10:29why it's important for the viewer to
10:31know this thing before we can move on to
10:34the next thing. The tension is where we
10:37teach the lesson essentially. We kind of
10:38reveal the information and then the
10:40payoff at the end of that um that
10:42segment is where we ultimately
10:45I suppose reveal the answer but like the
10:47the understanding of that segment is now
10:49complete and we can move on to the next
10:50one. And then the third one is a is a
10:52call to action. So whether that is a
10:53call to action at the end to promote
10:56your next video or product or service or
10:58a midvideo call to action um promoting
11:01whatever it might be a free lead magnet
11:02or or something like that. Can you give
11:04me an example hook on a video you've
11:06done for yourself recently? I like this.
11:07You said it was target transformation
11:09stakes. Target transformation stakes. I
11:12would love to hear like an example of
11:14that. Target is always very easy to me
11:16cuz I've got a five-page audience avatar
11:18document that I've made in in the way
11:20that we kind of described earlier. So, I
11:22don't need to think much about that. I
11:23know who I'm targeting. It's YouTubers
11:25who are interested in script writing. Uh
11:27script writers for hire who want to
11:29learn to get better at script writing so
11:30they can get more clients and blah blah
11:31blah. Transformation. If I take an
11:33example of one of my most recent videos
11:35and very uh apt for the conversation
11:37we're having now was how to set up a
11:39Claude project um for to to turn Claude
11:42into a reliable script writing partner
11:44uh as opposed to a sort of irritating um
11:47you know assistant who forgets
11:48everything. The transformation within
11:49that is is like you will go from not
11:53knowing how to do that to knowing how to
11:54do that. What's at stake is this idea
11:57and it's something that I think a lot of
11:59my audience have felt. Something I
12:00certainly felt when I was getting on
12:01board with AI is the sense that you will
12:04actually end up spending more time using
12:06AI to write your scripts than if you
12:08just use your brain. So so determined
12:10are you that this is the new thing and
12:12hey I should be more productive with
12:14this that you end up just getting
12:16frustrated with it. It doesn't quite do
12:17what you want and you end up wasting
12:18more time. And so for me that that's
12:20what's at stake if they don't achieve
12:22the transformation is you will remain in
12:24that state where this sort of perceived
12:26productivity of AI slows you down
12:28results in worse scripts uh means your
12:30channel doesn't grow kind of thing. So
12:32that's what I think before going into
12:36physically trying to write a script. And
12:37the joy of AI these days is that like
12:39you can say look here's my target here's
12:40my transformation here's what's at stake
12:42here is a hook framework that has
12:44performed really well on my channel in
12:46the past please express these factors in
12:50terms of this structure um and it's it's
12:53pretty good at very finite structural
12:56tasks like that a lot of the time and
12:58then out of that comes a perfect script
12:59and I'm good to go right that's it dude
13:01and I'm out of a job see you later
13:05yes so the joy way of all of this is
13:08that it this allows me to go from blank
13:10page to V1 bulletoint scripts very
13:14quickly, much more quickly than before.
13:16I'm hearing this and it's like I could
13:18do this while I'm out on a walk with a
13:20stroller. Like if I had this
13:21infrastructure set up um and I had an
13:24idea, I could literally do this while
13:26I'm out on a walk and it could create
13:28the artifact waiting for me when I get
13:30back to my desk. And that is a
13:32meaningful source of friction to
13:34eliminate that would make this feel much
13:36more real. And I would go from like I
13:38should make that video someday to like I
13:39bet I can record that tomorrow with a
13:41little bit of extra work here.
13:42>> Exactly that 100%. That's that's a
13:45really good point. Um, so not only does
13:47it save you that time in the moment sat
13:50down, it may well be that you don't even
13:51have to be sat down for it to for it to
13:52happen. But I think also it saves you so
13:55much mental strain having to restructure
13:59these thoughts on a page that
14:01perfectionism creeps in. It feels like
14:03even if you're just typing notes like
14:04no, I should arrange these in the
14:05correct order first and blah blah blah.
14:07All of that thinking uh power is is
14:10saved by just being able to talk to it
14:11like a like a person. So we have our
14:13structure say and we've agreed with it.
14:14Yes, maybe with a bit of back and forth,
14:16I am now happy with the six segments
14:19that we have for this video about Jay
14:21becoming a dad for the second time. Um,
14:24now I want to like move it forward into
14:25something more more word for word. This
14:27is kind of a bit of an inflection point.
14:29Maybe that's the wrong word for it. It's
14:30it's a a decision point on the the tree
14:34as to whether you want to say, look,
14:36this is as much help as I needed. Now, I
14:38sincerely want to be able to just do the
14:41rest of this by by hand, as it were, by
14:44brain. I don't want to then have an AI
14:46physically try and write me a full word
14:48for word script because I either that's
14:50not going to uh make me feel creatively
14:52fulfilled, or b I just don't trust that
14:53it's going to do what I want it to do
14:55yet. So, there is a point at which you
14:57might just decide this is where I finish
14:59the AI thing, and that's that's
15:00completely fine. If I were to say, I
15:02don't want AI to go any further, then I
15:04suppose you would then teach us how to
15:05think even more like a script writer and
15:07work through this, which is probably
15:09best to do anyway because the more
15:11knowledge we have, the the more
15:13collaborative we can be with bossing
15:16this machine around if that's what we
15:17choose to do. So, let's let's say AI
15:20turned around a V1 of the script and
15:22we're looking at it and we want to take
15:24this to from from this V1 to something
15:27that I'm ready to record. Let's say at
15:29this point we have our structure. I'm
15:31going to assume at this point we've
15:32mentioned bulletoint scripts a few
15:34times. It may be that you had your
15:36initial like here are the one sentence
15:38headline things for each of our
15:40segments. So like six literally six
15:42sentences. This is the skeleton. We're
15:44assuming we've now expanded it into uh
15:47say two or three bullet points per per
15:49segment. And so really the choice here
15:51is if you are someone who really enjoys
15:54uh riffing, if your niche is the sort of
15:56niche that will allow you to sit down
15:57and riff, then you can very much test
15:59that. You can think, okay, I'm literally
16:00going to go with what we have. Sit down,
16:02start talking, see how it feels in the
16:03moment. See how much pain is deferred to
16:06the edit because of how much I may have
16:07rambled if I was just going off of
16:09bullet points. And then adapt uh and
16:12adjust each time you sit down. You
16:13think, "Ah, actually, I felt completely
16:15out of control when I was sat down last
16:17time. This time I want to go from two to
16:19three bullet points per segment to five
16:22to six. I'm going to turn that dial up.
16:23I think of it like a dial. Just just
16:24turning it up a little bit. Testing it
16:26again next time. See how it feels. Did I
16:28ramble? Did that feel better? Was it now
16:30too prescriptive? Now I actually want to
16:31dial it back to four bullet points or
16:34something like that. So that's a really
16:35that's I suppose a personal thing. And
16:38that's if you choose the the riffing
16:39side of things. If you now want to go
16:43fully word for word, and this is
16:44something I massively prefer, the the
16:46amount of pain that is saved on editing,
16:49which is thankfully not even my job, but
16:51even just on filming day, the amount of
16:53false starts and thinking like, "Oh, did
16:54I screw this up? Wait, did I actually
16:56say that thing 5 minutes ago that sets
16:58up the thing I need to say now?" For me,
17:00that's just not worth it. I much prefer
17:02to put an extra 30 minutes, 60 minutes
17:04into writing a full word for word script
17:06to save me that pain later. If you
17:08prefer to spend less time scripting and
17:10more time recording, the OneUp of 10
17:12team just released a new feature that
17:13will help you edit your talking head
17:15videos fast. It's called auto cut and
17:17it's part of their new video editor. I
17:20have a great video setup here in my
17:22studio, but I don't actually like to
17:24script full videos. That means my
17:26recordings have all kinds of extra takes
17:28and dead space. And it's not that it's
17:30hard to edit those out, but it's
17:32timeconuming and I dread doing it.
17:34AutoCut takes your raw talking head
17:36footage, watches it frame by frame,
17:38removes silences, rambling, and retakes,
17:41enhances your audio, and gives you a
17:43clean cut that's ready to post in just
17:46minutes. It doesn't just put those extra
17:47takes in the trash, either. It saves
17:50them so you can instantly switch to
17:51whatever take you prefer. The One10 team
17:54is known for their deep knowledge of
17:55YouTube, and this tool is already better
17:57and faster than any other tool I've
17:59tried for cutting your raw footage
18:01quickly. visit the link in the
18:02description and you can try auto cut
18:04yourself for free. Then the focus and
18:08this is what I always focus on is
18:10obviously the hook is very important but
18:12the setups and the payoffs are the most
18:14important thing that I like to uh write
18:18word for word almost regardless. So your
18:20setup line at the start of each segment
18:22is is like another hook basically. It's
18:24the the the reminder of not only why I
18:27care about this overall topic, but why
18:29this particular subtopic is so essential
18:31in me moving forward. So if we take the
18:34example of my claude project video or
18:37yeah, you know how to how to set up a
18:38claw project. The first step of that was
18:42literally you need to set up a project
18:43because a lot of people don't do that.
18:45But the way we set that up compellingly
18:46is not step one set up a project on your
18:50Clawude uh your Claude chat.
18:52But before we get into the five
18:54documents you're going to need, we first
18:56have to fix one of the most common
18:59mistakes I still see dozens of my
19:01students making that turns AI into a a
19:04useless assistant or something like
19:06that.
19:06>> I see. I'm hearing that as an open loop.
19:08>> Yeah, exactly that. Exactly that. So,
19:10focusing on those open loops, those
19:12setups at the start of each segment and
19:13and just reminding the viewer in a
19:15curiosityinducing way why this next step
19:17is important. That's what those are
19:19those moments in your retention graph
19:21where people will go. It's when you
19:23finish one subject and you don't
19:25immediately make it clear why the next
19:27subject is going to be important to them
19:29specifically in the context of what they
19:31want and what the what the video title
19:33promised. If you had a literal script
19:35you've worked on in the recent past,
19:37especially if it's for your channel so
19:38you know that you have the permission to
19:39share it where you could be like, "Let
19:40me walk you through what the hook looked
19:43like in this video and then I'm going to
19:45show you one of these setups and payoffs
19:47in one of these segments." I think that
19:48would be enormously helpful.
19:49>> Can I do it on my app where I wrote it?
19:52>> Sure.
19:52>> This actually is that very video uh that
19:54we were just talking about. This is how
19:56to set up a claw project for script
19:58writing. So you can see this this is the
19:59software that I've built and so I break
20:01it up into these setup tension and
20:02payoff things that I've spoken about a
20:04lot. Very visually pleasing. This does
20:06make a lot of sense. It's modular.
20:08You're talking about the structure in a
20:09modular way. You're showing this in a
20:10modular way. And for what it's worth,
20:12anybody watching this, this is not a
20:14paid endorsement or partnership with
20:16George on this, but this is a great way
20:18to visualize it. And so, here we are.
20:20So, in terms of the hook, have you ever
20:21spent hours fighting with Claude? This
20:23already, even though it's not explicitly
20:24saying if you're a YouTuber who
20:26struggles with, but it's implicitly
20:28showing people, I understand you. This
20:30is the struggle that you have. So, if
20:32you've ever spent hours fighting with
20:33Claude only to check the clock and
20:34realize this YouTube script took two two
20:36times longer to write than if you just
20:37used your brain, then listen up. I'm
20:39using quite casual language in moments
20:41like this. With this video, we were
20:42experimenting with some like kind of
20:44comical intro B-roll. So, this is the
20:47idea that it's like I show that I
20:48understand you. I know who my target is.
20:50The problem with AI help scripts isn't
20:52that it's not good enough yet. It's
20:53because you're not giving it enough
20:54context. I guess this is kind of
20:56implying the stakes. It's a little bit
20:58tricky to always distill these into the
21:00exact three things, but like if I was
21:01listing these things out, I would be
21:04thinking when I list out my stakes, I'd
21:05be saying AI continuously takes longer.
21:08People don't understand anything about
21:10giving it context. They don't understand
21:12what is even necessary to improve the AI
21:14that they're using. And so for me, like
21:16this is how the stakes kind of filter
21:18through. But in any case, it's like this
21:20is a moment I often find comes up in my
21:22hooks is there's this kind of turning
21:24point moment. It's like a here's your
21:26current situation, but the truth is X.
21:29Well, the real problem is this. I've
21:30heard of a framework. It's P A I P S and
21:33the I is intrigue. And this is reminding
21:36me of intrigue. Like PA is pain,
21:39agitate. We've we've kind of talked
21:40about that already, but intrigue is like
21:43sort of a point of surprise before
21:45introducing a potential future as that
21:47next P. So, this is this line makes me
21:50feel like ah it's a it's a head turner.
21:53It's like you know something that I
21:54don't know.
21:55>> Yes. Yeah. Yeah. For sure. Having
21:56written the last eight videos with the
21:58help of AI. Again, this is we're kind of
22:00talking about that future state, but I'm
22:01not explicitly being yet like if you
22:04want to write your your next uh your
22:06next script using AI. It's sort of both
22:08implying my own credibility, but also
22:11it's uh it's presenting an image of
22:13their future state. So again, I think of
22:14that as the transformation expressed
22:16through credibility, I suppose. And
22:18spent the last 3 months in the trenches
22:20training my own script writing app. I
22:21think I got rid of in the trenches in
22:22the end cuz when I spoke it just sounded
22:24a bit weird. I'll show you the exact
22:26five documents I think it became in the
22:28end.
22:30Uh you need to turn Claude from a
22:31forgetful assistant into a reliable
22:32script writing partner. Uh so that's
22:34fully the transformation that we're
22:35ending on. So that's how I think about
22:37the hooks. And then now we come down to
22:39the first segment. So we're not saying
22:41in this first segment yes that it's it's
22:43about making a new project but we're not
22:44saying the first step is making a
22:46project. So that's how I set it up.
22:48Tension is a really interesting one. We
22:50could talk for quite a long time about
22:52uh about tension, but the way I like to
22:54think of this is, if I back up for a
22:57second, the problem people have with
22:59payoffs is they feel like they have to
23:01defer giving any value at all until the
23:03payoff. Like it feels like it has to be
23:05this big explosive moment at the end of
23:06a segment,
23:07>> right? Like one payoff per video. I'm
23:09thinking about this like these are like
23:10you're you're uh nestling multiple
23:13payoffs into this. The whole video
23:15itself is a payoff. There are five
23:17payoffs that add up to the payoff is
23:20what I'm seeing. Exactly. Exactly that.
23:23Um, so yeah, lots lots of payoffs
23:25throughout. But even within that, I
23:27still see people thinking that like,
23:29sure, I get the idea I have to have a
23:30payoff at the end of my segment, but
23:33that means I have to literally say
23:34nothing of any value until I hit it, and
23:36the rest is all kind of preamble. What I
23:38like to do is think of it in this
23:39slightly different way where it's like I
23:42start by showing people kind of
23:45I suppose we've done a little bit of
23:46this in the hook, but we're now going
23:47into detail on the user's present state.
23:51So it's once again showing at the start
23:52of this that we we've got their
23:53curiosity with the setup. Now we're
23:55showing here is the thing you probably
23:57do currently. So, in this case, it's I
23:59was surprised to find a bunch of my
24:00students were still using Claude in an
24:01extremely siloed off way, starting a new
24:03chat every time, explaining who they
24:04were and what uh what they wanted AI to
24:06help them with. This is very much what
24:08we've been discussing in this podcast
24:09anyway. And then we kind of move through
24:11into here's what you actually need to
24:12do, create a project, here's what a
24:14project is, um, and then go and create
24:18it. This is a slightly atypical one
24:20because this is so step by step. Go and
24:22do a specific thing. It's normally a bit
24:24more kind of nebulous, but for anyone
24:25who does have a channel where you're
24:26showing people step-by-step things, I
24:28think this would be the key takeaway
24:30here is make sure that you rather than
24:31just saying step one, do this. It's
24:34explaining in a cur opening a loop about
24:36why this is important, then showing them
24:39what they currently do and why it's bad,
24:41why it doesn't work, and then you can
24:43move through into here's what you should
24:44do instead. So then the payoff line and
24:46this is I've kind of shifted the way I
24:47think about these over time but you'll
24:49notice that it's not really like
24:52we're giving the final like here's what
24:54you have to do in that payoff line. It's
24:56more bringing people to a level playing
24:58field and saying so here is the state we
25:00are now in as a result of what you have
25:02just learned. Here is the new status
25:04quo. You have now done that you are set
25:06up and we are ready to do whatever comes
25:08next. It it almost reads like it's a
25:10setup for the next setup which makes me
25:12think about our first conversation where
25:13we talked about signposting. Is this how
25:16you see signposting today? To a degree I
25:19think maybe this is a slightly atypical
25:20example because we still want the next
25:22setup to be curiosity inducing and open
25:26a loop which is why you can see here
25:27it's like I say here the first document
25:29I'm not going to say the first document
25:31is this. It's the first document is
25:33going to help you fix thing that you're
25:35curious about. But if I look at
25:40any other examples in here that that
25:41explain this? Well, I think this is
25:43probably a good example that that shows
25:45it off. I I wouldn't say we're
25:48necessarily signposting what comes next
25:50yet. Um I do think that is reserved for
25:53the following setup, but again, it's
25:56like clarifying with the information you
25:59didn't have that you now have that you
26:00learned in this segment.
26:02Let's just take a pause to reflect on
26:04like what what is it that we have now
26:07unlocked the ability to do? Where is it
26:08that we now stand in the overall context
26:11of this topic? And I think this is a
26:13good example of this where after we've
26:14uploaded, this is I think the third
26:16document that I show people. We're now
26:18just clarifying. Okay, take a pause.
26:20We've been saying all this stuff about
26:21these documents you have to upload, but
26:23practically speaking, let's zoom out.
26:25What can your project now do so that you
26:27understand why we're doing all of this?
26:29Yeah, tell me if this resonates with
26:30you. I think I'm following what you're
26:31saying, but the other thing that I'm
26:34latching on to is some speaking advice
26:36I've gotten, which is tell them what
26:37you're going to tell them. Tell them
26:39tell them what you told them. And this
26:41almost mirrors that a little bit in some
26:43of these cases, but maybe maybe the
26:45payoff isn't like totally saying, "Let
26:47me close the loop on what I just told
26:48you, but it seems like maybe it could
26:50be."
26:51>> Yeah. I This is a really
26:54kind of tricky semantic dichotomy.
26:58Sounds like the name of a band.
27:02Yes, I think I agree with you in
27:03principle, but I think we want to try
27:04and avoid it's treading the line between
27:06that and not just making it a summary.
27:08So, yes, tell them what you told them,
27:10but I think I want to do it in a
27:12slightly different way. So, I'm not just
27:14repeating, so you just added this
27:16document. It's more like um you added so
27:19having added this document, new
27:21information, this is now what your
27:23project is capable of thanks to what you
27:25just did. Um rather than just simply
27:28repeating it.
27:29>> Yeah. Does that distinction make sense?
27:30It's kind of a tricky one.
27:31>> And that's kind of what I'm trying to do
27:32with like these comparisons cuz when you
27:34tell me like yes, but it it clarifies
27:36more like what is a payoff then? What is
27:38a good payoff? It's not a summary. It's
27:41not a setup for the setup. Uh it's
27:43putting a period on the end of the
27:44sentence. Yes. But it's also it's a
27:46breath in one way. It also seems like
27:49it's making the person feel good and
27:52excited and accomplished in another way.
27:55Uh which obviously summaries don't do.
27:58Yes. Exact. I think that's a really a
27:59really good distinction. Um, make people
28:01feel good. Catch them up as it were.
28:04Like if anyone's kind of gotten lost or
28:05they're wondering, wait, have I have I
28:07grasped this? Just taking that moment to
28:09say, yes, you've got this and doesn't it
28:10feel great? Um, as opposed to saying,
28:13here's what you just learned. In much
28:15the same way, the AI in V1 is going to
28:18give us like, here's how many sections I
28:20think we should have. Is the goal in
28:22your mind to reduce the number of
28:25sections
28:27necessary? How do I think about the
28:29number of sections that are in a video?
28:31I always come back to the the wise man
28:34Gandalf uh when I'm asked this question,
28:36which is that a wizard arrives precisely
28:38when he means to. Uh this is a tricky
28:40one. There are kind of no rules to this.
28:42I think a lot of it comes from just the
28:43soft skill of doing this over and over
28:46again and figuring out, okay, when I
28:50talk about X in my videos, my retention
28:53always drops off. On reflection, this
28:55period where my retention was kind of
28:56diving continuously, I realized that I
28:58was sort of talking about the same topic
29:01but in slightly different words twice in
29:03a row or something like that. I think
29:05really it just comes down to honest
29:06reflection and looking back at what
29:07you've written. Um, I I know one thing
29:10that can certainly help with this is
29:11just ju just showing it to someone else,
29:13whether it's in your niche or even
29:14better sometimes someone who's not in
29:16your niche. Someone who can look at it
29:17objectively with no um feeling that or
29:21no investment in the topic, no kind of
29:23preconceptions of what it's about and
29:25just saying like is this interesting to
29:26to read? Did did it fall off at any
29:29point for you? When did you get bored?
29:30So outsourcing it is a is a good one.
29:32But I think overall, if we're just
29:35talking you sat at your desk, I think
29:37having no preconceptions of how many
29:40segments a video should have uh is key.
29:43If you're trying to target a certain
29:45length of your video, I know people like
29:48to try and have longer videos for the
29:49watch time, whatever. But if you've got
29:51an uh overly inflated video that doesn't
29:53have any substance to it, then it's it's
29:55not going to be interesting. So, I would
29:56say a wizard arrives precisely when he
29:58means to. How long is a piece of string?
29:59Like, make it the number of segments
30:00that it has to be. be objective and look
30:02at am I including this segment because
30:05it contributes to the understanding of
30:07the subject or am I doing it just to
30:09bloat the watch time. Something that
30:11I've been trying to push myself to do
30:12more of is just slow down and feel
30:16better about every step of the process
30:19because that's the input I can control
30:21that will probably lead to better
30:22outputs. But I get to this point where
30:23it's like I'll get really excited when I
30:25feel like I have the idea that I want
30:27and it's like I'm just going to go rip
30:28that off and record that. and
30:31underinvesting in pre-production I think
30:34is like the the biggest tax I'm paying
30:38right now on the channel is is I'm just
30:40underinvesting in pre-production all the
30:43ways and you see that cost in the
30:47performance of a video. If a video
30:49doesn't do well and you you think well
30:51like well I was proud of it and I did
30:54everything I can. that's different than
30:56I should have taken more time on that
30:58because then you're probably not going
31:00to go and just remake that same video
31:01the next week. It's it's just worth it's
31:05worth feeling better in pre-production.
31:07One one just thought that came to me as
31:09you were saying that there is I think a
31:11a way of making sure that you're still
31:13putting some of that thought and time
31:14into pre-production even if you're short
31:16of time especially with AI now is even
31:19if you feel like you don't know what you
31:22want your video to be and you don't know
31:23how you want to structure it and you
31:24feel overwhelmed by the idea of it
31:26spieling to an AI and having it present
31:28a structure back to you will immediately
31:32set off red flags in your mind of things
31:34that it definitely got wrong. So if you
31:36look at a structure that that it has
31:38come back with I I I guarantee and maybe
31:42it'll nail it but most of the time it
31:44will have come up with some idea that
31:45you're like no no that's totally not it
31:46and then you realize oh actually I do
31:48have some thoughts on this and I do have
31:49my opinions on this but having it
31:51presented to me in like slightly the
31:52wrong order is much easier then for me
31:55to think okay no I would shift segment
31:56one to segment two I would maybe add an
31:58extra segment about this I would connect
31:59these two ideas like this and then
32:01suddenly you've got your your structure
32:03out and it feels like you have invested
32:04that time in pre-production.
32:06Um, but you're sort of spurred into it
32:08by seeing what was definitely not the
32:10right structure. Um, as opposed to just
32:13trying to do it from a blank page.
32:14>> Yeah. Can you go back to sharing your
32:16screen? I want to see you talk about
32:17your CTA now.
32:18>> Yes. Okay. So, um, in this video I
32:21didn't have any product CTAs. I I think
32:26because I mentioned my AI script writing
32:28tool in the hook, I just put the first
32:30link or the second link as a link to to
32:33the weight list for this tool. That's
32:34actually been a really effective call to
32:36action for me is is not mentioning it at
32:38all, but just showing the thing that I'm
32:40talking about the whole video and then
32:42putting a link in there um in the
32:43description. People intuitively just
32:45want to go and and learn more about it.
32:46>> And in that case then, is there no end
32:50besides like maybe an end card like you
32:52talked about in our first conversation
32:54of just like tying it to a next video or
32:56do you just let the payoff of the final
32:58segment be the the end of the video?
33:00>> No. So this is my absolute every single
33:02time without fail thing that I do is
33:05every video has to link to another one
33:07in a in a compelling way. But at the end
33:10of every single video I'm thinking about
33:11a link. So that is uh back to something
33:15in this video. So we're trying to make
33:18it not apparent that the video is about
33:19to end. We're basically saying we've
33:21already spoken about XY Z thing or now
33:22you know how to do blah blah blah blah
33:24blah. If you couldn't see the timeline
33:25at the bottom, you might not know that
33:27the video is about to end. Then we want
33:29to add some uh curiosity. So um about
33:34something in the next video.
33:37Um so in this case here, if we go
33:39through these one by one, link back to
33:40something in this video. Now this
33:41project setup is a great starting point.
33:43Done. It's like simple as that. That's
33:44all you need. Um that shows that we are
33:48talking about we're referring
33:49hell. Sorry. Grammarly on my
33:52>> I will leave when I decide.
33:54>> Yeah, dude.
33:55>> It's okay. It's okay. We can read it.
33:57It's fine. I'm just removing you from
33:59Chrome. Grammarly, you are you are so
34:00gone. Um, no, it's still here.
34:06This is ridiculous. Let's see. I laughed
34:09so hard that I unplugged my own
34:10headphones.
34:13Okay, so
34:16if we take a look here, this is clearly
34:18the link. This is where we say, now this
34:20project setup is a great starting point.
34:21It's just referencing something uh well,
34:24basically the entirety of the topic of
34:26this video.
34:27Then we raise the curiosity here where I
34:29say, "But what if I told you we can ask
34:30Claude to gradually improve itself over
34:32time?" So each new script you write gets
34:34closer and closer to your style every
34:36time. I think in the moment I must have
34:37rifted something else here because I say
34:39time twice in quick succession. Uh we
34:41want to avoid repetition. And then
34:43finally, we just have the promise uh
34:46which is usually spelled like this.
34:47There we are. uh CTA
34:52which uh is where we give the explicit
34:54promise and this is where the facade is
34:56kind of broken and we're like we yeah
34:57we're promoting another video go watch
34:59it but ultimately it's an offer right
35:01we're giving the viewer something else
35:02for free but watch this video next where
35:04I'll walk you through how to do this
35:05using a simple prompt that you already
35:07have inside your claw product uh project
35:09so it's like we're promising something
35:11that is exciting for you and feels low
35:13lift so this is like your non-negotiable
35:15this ends every video I have a question
35:16on this point because I saw an
35:19interesting strategy and I'd love to get
35:20your take on this cuz I think there's a
35:22cost to the strategy that you're going
35:24to say as well, but I wonder if the
35:25benefit is is worth it. So, I saw a guy
35:28who had um an unlisted video on his
35:33channel. This is talking more. He had an
35:35unlisted video that's essentially a
35:36video sales letter. And what he does is
35:40he ends some of his videos, not every
35:42video, but a video that like clearly
35:44showed some credibility. and somebody
35:46who watched the end of that video is
35:48probably a qualified lead for his
35:51business and his end card and his end of
35:55the video is linking to his unlisted
35:58essentially VSSL. What do you think
36:00about that strategy? I think that's
36:01really cool. I I love that. I think like
36:04anything, it's kind of balance and
36:05you're trying to figure out what am I
36:06optimizing for mostly here. Um, the
36:09advantage of of this strategy and always
36:10ending with something else is that you
36:12can create these kind of infinite loops
36:13of people just watching more and more
36:14and more stuff. And it doesn't matter
36:16which video they land on first, there's
36:18always going to be kind of infinite
36:20trailer videos they could keep binging
36:21if they wanted to. I suppose the only
36:23risk with this strategy is, hey, if
36:25someone if it's like the first video
36:26that someone's landed on is that one and
36:28they get pitched something at the end,
36:30what is the effect that has on them? Do
36:31they think, oh, I've been sort of sold
36:33to here. Um, it was kind of an ask, not
36:35an offer. But ultimately, like I I think
36:38that sounds like a really a really good
36:39tactic and one that I might steal. Um
36:41and clearly you would think he would
36:42have been testing something like that,
36:44right? And the the pros would outweigh
36:45the cons.
36:46>> Yeah. And I imagine that same video
36:47could probably be like a good channel
36:48trailer. I know Ed from Film Booth has
36:50tried stuff like this where like his
36:51channel trailer is essentially a VSL. I
36:54think it's really smart. I mean the
36:55biggest pain point we have on the
36:56channel and candidly why we've had to
36:57pull back some resources on the channel
36:59was we were spending far more than we
37:01were clearly earning from the channel
37:04because we never invested in connecting
37:07the channel to the core business and
37:10this is like a really consistent way to
37:13know that this is working if you have a
37:15video that you know when people watch
37:17this video they're interested in
37:18learning more and going deeper with me
37:19and if that is true then I think this
37:21actually makes the YouTube as a
37:23discovery engine for the business a much
37:25shorter less uh luckfilled
37:30like uh process for getting people into
37:33the business. Yeah, I think um again
37:35that's something you can really easily
37:36start using AI to help you with where
37:38it's for something like this that has a
37:40formula that I try and use anything
37:42that's formula based. If you've got
37:43something simple, we spoke about it with
37:44the hook as well. It's very easy then to
37:46plug in the
37:49actual content that you want and to ask
37:50it to express that content in terms of
37:52that formula. So in your case, I suppose
37:54using a formula like that, but saying
37:55like, "Hey, actually, we don't want to
37:57promote another video. We want to
37:58promote X or Y product. We want to
37:59promote the lab. We want to talk about
38:01whatever. We want to promote a VSSL that
38:03talks about the lab or something." I
38:04think it would be fairly easy then for
38:05it to come up with a relatively subtle
38:08plug in the the style that we've just
38:10seen here, but for a VSSL. And it's a
38:13spectrum because it doesn't have to be
38:14like a straight up like sales video. It
38:17could be a video that you just know when
38:19people watch this video, a lot of them
38:20turn into customers. So, how can I feed
38:23more people into that video even beyond
38:25its own packaging and and CTR? I think
38:27is a a smart strategy. Intentionally
38:30antagonistic question. Okay.
38:34>> Yeah. Because you are a script writer
38:37and probably a lot of people who have
38:38historically been like, I could be a
38:40freelance script writer. I'm going to
38:42get into this. there is a certain amount
38:44of like incentive
38:47if not bias towards saying like we
38:51should keep human script writers. So if
38:53you were to take as much of an objective
38:56stance away from yourself as you can
38:59where would you say the human input the
39:03human in the loop in this process is
39:05most essential right now? That's such a
39:07good question. I think to aid my case
39:10here, at least in terms of the
39:11objectivity, I'm in the fortunate
39:13position where I've managed to sort of
39:14work my way out of having to write
39:16scripts for people. And so I think I'm
39:19now able to grow my personal brand or
39:22whatever you want to call it as someone
39:24who helps people learn to do script
39:25writing in whatever context that is. And
39:27now that context is AI related. So, I
39:29think I'm able to divorce myself a
39:32little bit from the desire to like have
39:33script writers around because it's what
39:35I used to do a lot and it's um it's what
39:36I still do for my own channel. One thing
39:38that I've learned a lot about since
39:39starting to really use AI a lot and
39:41using it to code my own script writing
39:43app, for example, is I've spent a lot of
39:45time watching how AI actually does what
39:49it does behind the scenes. Like you know
39:50in Claude for example you can sort of
39:52see how it's thinking as it's thinking
39:53and you can read its decision-m uh you
39:56sort of learn all these little ways in
39:57which it may appear to be doing
39:59something well and then the output that
40:01comes out it actually reveals oh hang on
40:02it wasn't doing these like 10 things
40:04that it claimed very confidently that it
40:05was. So I think when it comes to script
40:08writing the final say on like when have
40:11I done enough revisions for this script
40:13when is this actually record ready that
40:16ultimate decision comes down to a
40:18person's brain because AI will will
40:20always have something else that it's
40:21like oh we could change this and oh you
40:22actually didn't do this correctly and oh
40:24this could be stronger so we could make
40:26it land better or whatever and it will
40:28just keep doing that kind of infinitely
40:30and I think that moment where you have
40:32to say to yourself no actually I think
40:34this is ready I know my audience
40:36as wellroundedly as I possibly could.
40:39I'm the one who says that we've now done
40:41enough revisions on this and I think
40:42it's going to it's going to hit really
40:43well. But I think there are those
40:45moments that come throughout the
40:46process. So it might be when structuring
40:47out the script in the first place. AI
40:49will be like, hey, here's a brilliant
40:51structure based on your audience avatar
40:52that I think they're going to love. And
40:54you can read that and be like, I like
40:5670% of these ideas, but ultimately I
40:58want to rearrange X and Y because I know
41:00this will be how I would actually
41:01explain it or how they would receive it
41:03better. And so again, the point of
41:05deciding structuring is now done. That
41:07kind of needs to be me because AI will
41:09keep on coming up with new things that
41:11you should add or change or improve.
41:13there is a certain courage to
41:16cut the AI off, you know, and I think
41:19that's that's kind of a meta-kill
41:21through all AI usage that people need is
41:25courage and trust in your own
41:28intuition, your own taste, your own gut
41:32to know when
41:34you're getting into like the recursive
41:35improvement for improvement sake that is
41:38no longer actually improvement. Maybe
41:41this is related, maybe it's not, but
41:42it's an aside I want to take. I I have a
41:45similar video that I want to make, but
41:47it's it's kind of talking about like,
41:48hey, I'm about to have a second kid, and
41:51the first time I did this, I didn't
41:53prepare well, and so I'm preparing
41:54differently this time. And where I kind
41:56of get hung up is I don't know if
42:00there's enough there. There when you're
42:02kind of laying this out through I'm
42:04imagine you're using like whisper flow
42:06or some sort of dictation. When you're
42:08laying this out, when do you feel
42:10confident that you've got a video that's
42:13long enough to warrant existing? Oo,
42:16that's a really good question. I think
42:18with a lot of anything creative, there's
42:21an element of brutality that needs to
42:23come that comes down to it in the end.
42:26There have been some scripts that I have
42:27written which have seemed like a great
42:29idea have been maybe like five segments
42:32or kind of subsections initially and
42:34then as we've kind of workshopped either
42:36me personally or me with in
42:37collaboration with an AI have
42:39workshopped it it's kind of trimmed down
42:40and down to maybe the point where
42:42there's like three segments and it
42:43reaches a point where I'm like I think
42:45this could just be a newsletter or
42:46something you know that's a bit more
42:47short form and so I think the first
42:49thing would be to say accepting the
42:51reality that it may be that there just
42:54isn't enough to the topic but I I think
42:56this is one of the real meta skills to
42:59to to learn and it's something I'm sure
43:01you're you're aware of anyway is like
43:03just the ability to talk to AI to
43:06extract from it exactly what you want.
43:08So talking to it in natural language and
43:09say look here's the predicament I'm in.
43:11I have this exact topic I want to talk
43:12about. Here's the context on my last
43:14video about having a kid. Uh this one I
43:16want to do now is about how I wasn't
43:17prepared properly but I'm worried there
43:19isn't going to be enough to talk about
43:21it. Knowing what you know about my
43:23audience, if you don't have an audience
43:25avatar, like here are 10 transcripts of
43:27some videos we did recently. Here are
43:29some recent comments from them. Knowing
43:30what you know about them, what blind
43:33spots do I have that could relate to
43:36this topic that we could put into this
43:38video? And kind of using it then to
43:39challenge you and and to almost make you
43:41answer some questions you might not have
43:43even thought to ask yourself that can
43:44then contribute to uh making a video
43:47that feels like it warrants existence.
43:49Uh to kind of close us out here, what's
43:51your relationship to AI right now?
43:53Obviously, you're embracing it for
43:54making a tool and you're using it in the
43:55script writing process. Are you just
43:57like full in like, "Yep, toothpaste out
43:59of the tube. I'm using it. I love it."
44:01How do you feel about AI? It It's an
44:03interesting moment, isn't it? Because
44:05it's quite It becomes harder and harder
44:07to extricate yourself from it and to
44:10reason with yourself and think as you
44:13would have thought before what's now
44:15possible
44:16were possible. my business didn't rely
44:19on me moving with the times like this,
44:22would I be going as all in on it as I
44:24am? I don't know. Uh I'm certainly
44:27loving building an app. This this is
44:29something that I had spoken to different
44:31companies about in the last few years
44:33that I sort of being brought on as like
44:35a
44:38as the sort of the the brains as it were
44:40as the script writer behind a tool that
44:43would then be coded by a bunch of very
44:45intelligent people. the fact that those
44:47very intelligent people are now not
44:48strictly needed. Although I'm sure
44:50anyway there's a whole discussion there,
44:51but it's I I love that I've been able to
44:53build my own thing. This is the tool
44:55that I've dreamt of making and it does
44:56all the things that I've always wanted.
44:57That's very cool.
45:00Do I feel the most creatively fulfilled
45:02that I have uh in the four or five years
45:06I've been running this business?
45:08In some ways, no. And as time goes on,
45:12the more I kind of want to shut it all
45:14down, go live in the woods, and makeong
45:16YouTube videos just chopping wood. This
45:18is exactly how I feel. I feel like the
45:20more I embrace AI, the less I enjoy my
45:22work. And I like I myself start to feel
45:25like a factory rather than any type of
45:28artisan. like anything that I've used AI
45:32to do to get to the point of done and
45:34skipped the doing
45:37>> I just have no affinity or pride in that
45:42work. Uh not to say it's bad but to say
45:44like I just don't feel pride in it
45:47because I didn't do it.
45:48>> This is my biggest thing is like
45:50>> okay this can help me do a lot of things
45:52that I have been doing but
45:55to what end? I've got to find the things
45:59that are absolutely non-negotiable that
46:01I enjoy taking pride in that I want to
46:02do. And for me, writing like the essays
46:06has been part of that. Writing the book
46:08is part of that. YouTube scripts, like
46:10we we had a video we did and I just I
46:13don't get excited about writing the
46:15intros for our podcast episodes. And so
46:17I was like, here my AI agent, here are
46:20my top 10 performing videos. Learn what
46:23you can about the structure. Here's the
46:25transcript of the rest of the edit. Help
46:28me make an intro for this. And it did as
46:32good of a job as I would have given my
46:35current skill set and the time I would
46:37put into it. Did I then show that off
46:39and say, "Look at what I made uh for
46:41that intro?" Absolutely not. Um so it's
46:44I'm I'm deeply conflicted. Yeah, I think
46:47I I'm completely with you. It's hard and
46:50I don't know if you feel this especially
46:51from a business point of view where I'm
46:54I'm trying to build an AI related
46:57business that is
46:59that that is um that doesn't completely
47:02remove the human element. I don't want
47:04to build a script writing tool that
47:05writes every single word for you and you
47:08go away and you record something that
47:09you just didn't you had no nothing to do
47:11with. which is why the sort of core
47:14feature of it is this discursive element
47:15that we've discussed where it's like
47:17there's a there's a chat you chat to it
47:19you leave your voice notes and you kind
47:21of work with them as if they were a
47:23script writing a script writer that
47:24you'd hired but at the same time I'm
47:26very conscious that this tool also would
47:28completely serve people who want to
47:31start a YouTube channel as a cash cow
47:35faceless no shade on faceless in general
47:37but like one of those channels where
47:38it's like hey I don't have to do
47:39anything I can AI the voiceover this
47:41guy's got this tool out and it can just
47:43AI the script and I can AI the
47:46thumbnail. And so being being conscious
47:47that you're kind of you're serving both
47:49the more conscious people who want to
47:52keep the human element involved but also
47:54the people who want to just outsource it
47:56all. And yeah, that is kind of
47:58conflicting. It's a weird a weird
48:00moment.
48:00>> Yeah, it's tough. I think um I think
48:02people want to maximize their outcome
48:07and minimize their effort and you follow
48:10that all the way down and people want
48:12everything just like fully out of the
48:13box done. When you actually watch their
48:15behavior, that's what they want most of
48:16the time, most people. So, it's yeah,
48:19it's it's challenging and it's I
48:22restrain myself from the temptation to
48:24become like an AI educator because
48:27there's so much tailwind to it and it's
48:29what people want, but it's just not what
48:31I want to do. It's an interesting time,
48:34but I appreciate you uh sharing with us.
48:36We'll put um a link to the the document
48:39you recommended and this tool that you
48:42showed us in the description. George,
48:44thanks for coming back on the

About this transcript

This page contains the full transcript of His EXACT System For Writing GREAT YouTube Scripts With AI by Jay Clouse, generated from the public captions YouTube serves with the video. The transcript has 10,620 words across 1,441 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.