ChatGPT 6 Astra & Claude Gave My Channel An Unfair Advantage
A YouTuber walks through the phone-first AI workflow he credits for turning his channel around: one AI project per video, an audience document, and talking points scored by viewer interest.
A channel grows when the AI is guided by documented audience interest, so titles, thumbnails and talking points come from what viewers already respond to rather than the AI's default idea of a good story.
Who This Is For
Read if. Skip if.
READ IF YOU ARE…
A small YouTuber with a few videos or an audience that has stalled and who wants a repeatable pre-production process that fits on a phone.
A faceless or voiceover creator using AI voice who needs a way to organize talking points around what viewers want.
A creator who already uses ChatGPT or Claude but gets generic outlines, clever-but-unclear thumbnails and weak retention from them.
SKIP IF…
You want a tutorial on AI video editing or motion graphics. The video says that is not the focus.
You want hard analytics proof. The metrics shown are a handful of the creator's own videos.
TL;DR
The full version, fast.
Brian G. Johnson says his channel turned around once he stopped using AI out of the box and started guiding it. Each video gets its own ChatGPT or Claude project loaded with documents about his audience's wants and struggles. He voice-dictates ideas, and the AI returns talking points scored by viewer interest, so he can follow a structure that keeps people watching. Titles start from outlier videos in his space, thumbnails are validated against written rules, and the best points go first, not last. He targets a click-through rate near 7% and five-plus minutes of average view duration.
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.
Results shown, a phone and an AI plan are all you need, and an open mind is requested.
01:58 – 03:35
02 · The scored outline
On-screen outline with numbered points and audience-interest scores that show what comes next.
03:35 – 06:57
03 · One project per video
Voice-dictate ideas into an AI project loaded with audience documentation and have it organize talking points.
06:57 – 10:26
04 · Outliers set the title
Watching what wins on YouTube each day and starting from a proven title format.
10:26 – 14:51
05 · Why browse rewards it
YouTube's own description of how browse works explains why topic overlap earns impressions.
14:51 – 16:45
06 · Not every video is an outlier
Examples of videos built from audience questions that still performed.
16:45 – 20:43
07 · Thumbnails and editing
Validating thumbnail ideas against documented rules, and letting AI handle grunt-work editing.
20:43 – 22:30
08 · Sponsor: 1 of 10
A tool for spotting outlier videos and tracking channels in the creator's space.
22:30 – 26:00
09 · Where the system fails
AI saves the best points for the end; lead with them and target a 7% CTR with five-plus minutes watched.
Atomic Insights
Lines worth screenshotting.
The structure keeps people watching, not the script itself, so a scored outline can replace a written script.
Give every video its own AI project and load it with audience documentation so the AI guides itself by your viewers instead of generic best practice.
YouTube's browse system weighs how often a viewer watched your channel and how often they watched the topic, so a new channel borrows reach from the topic.
Starting a video from a title that already won as an outlier makes the final title far more likely to be a winner.
A contrarian thumbnail pattern that worked years ago still works because people respond to the same triggers.
As a channel uploads more, viewers who already like it click without scrutiny, which buys freedom to cover topics that are not outliers.
A thumbnail needs clarity, not beauty: a short phrase of four words or fewer that matches what the audience desires.
Tell the AI not to give a thumbnail concept until it has validated the idea against your documentation.
AI defaults to saving the best payoff for the end, a spot only about 20% of viewers reach.
Putting the big payoff within the first 17 seconds gives viewers a quick win and extends average view duration.
Click-through rate falls as a video reaches colder audiences, so retention has to carry the video once CTR drops.
AI is a poor thought partner for structure and will damage it unless you supply the viewer-interest rules.
Takeaway
Guide the AI with your audience, not defaults
WHAT TO LEARN
AI helps a channel only when it works from written audience interest, which shapes the title, thumbnail and the order of points.
02The scored outline
A visible outline with numbered, scored points makes it easy to follow a structure that performs, so a full script becomes unnecessary.
03One project per video
Keep a separate AI project per video and load it with documents on your audience's wants, struggles and what has already worked on your channel.
Dictate your ideas by voice, then have the AI order the talking points by viewer interest so you can follow a structure that holds attention.
04Outliers set the title
Begin every video with a working title informed by outliers in your space, since a proven title format raises the odds of clicks.
05Why browse rewards it
YouTube shows videos to viewers who watched similar topics, so overlapping with winning videos in your own style earns impressions from people you do not yet have.
06Not every video is an outlier
Videos that are not outliers still work once viewers have watched your channel several times, which frees you to cover your own topics.
07Thumbnails and editing
Judge thumbnails on clarity: four words or fewer, instantly about your subject, tied to what the audience desires, and validated against your rules first.
09Where the system fails
Lead with your strongest point in the first seconds, since AI tends to put the payoff at the end where few viewers arrive.
Aim for early retention and a decent CTR together, because a falling CTR on a colder audience must be offset by longer watch time.
Glossary
Terms worth knowing.
Outlier
A video that performs far above its channel's usual views, used as evidence that a topic and title format resonate with an audience.
CTR
Click-through rate: the share of people shown a thumbnail and title who click to watch.
Average view duration
The average amount of time viewers spend watching a video, a key signal for how well it holds attention.
Packaging
The title, topic and thumbnail of a video, which together decide whether someone clicks.
AI project
A feature in ChatGPT and Claude that keeps a set of uploaded documents and instructions attached to a conversation.
“What happens if you give that big payoff at 17 seconds into the video?”
concrete number and a clear lesson→ newsletter pull-quote↗ Tweet quote
The Script
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.
17px
Hey, I'm glad you made it. So ChatGPT and or Claude have given this channel a massive advantage. Almost like an unfair advantage.
In fact, if you look at the results, you'll see some of the videos, one in particular, 100 ,000 views, 41 ,000 views, 36 ,000 views, and on and on. Not so much that the AI is doing a lot of the whiz -bang, editing a video, creating a motion graphic. Is that what you're interested in.
I don't think so, but I don't know. But based on my wants and my desires, I want more views. I want success.
And I want to walk you through this system that is so stupidly simple. Anybody can do this. AI, talking head, maybe your own voice, but you're not on camera.
This can be implemented by anybody and it's... It's crazy. Now, this video I published just a day and a half ago.
This is the craziest. It's one of 10. It's driven around 6 ,000 views.
The metrics are not like that impressive, yet the system is ensuring the stuff that needs to happen happens and it's working. So the average view duration is like three and a half minutes. And the CTR is just maybe a little under 7%.
I want to ask you just to have an open mind, and if you want to succeed, one idea has helped me, and that is, I wonder how I could implement this. That's it. And I think if you approach this video with, hi, I wonder how I could implement this.
This is what you'll need, and I'm already getting off of my script, but you'll need a phone and an AI plan, and I think you could probably even do this for free. or really take it far, what I would do is use a cheaper model that has, it's not as expensive, obviously. And that way you can push it farther.
And we'll get into the how -to, the nitty -gritty. Oh, and by the way, if you're new to the channel, my name is Brian. Let's jump in.
Now, what you're looking at right now is the actual outline of the video. And I'm just, I've got my script. This is the script.
I don't think anybody should script anymore. It doesn't make sense. That's another video.
But this is the outline. And the outline does a bunch of things to ensure I'm keeping those people watching as long as possible. So hopefully you find this engaging.
Not everybody will, and that's okay. But the outline, it's interesting in that you can see what comes next. So right now I'm on number six.
You can see what comes next. And notice the score. 96.
Now, the next talking point is number seven, and it says, notice the numbers besides besides each point. So again, it's really easy for me to follow a structure that leads to better performance, which is kind of why scripting is so powerful.
The idea of a script is it's not the script, it's the structure that keeps people watching. It's the structure in knowing that the first 30 seconds, right, we've called that a hook for a long time, well, the hook is pretty solid. proof it's working all these things and there's always some way to create a structure no matter where you're at just getting started failing upload succeeding there's always a way to to kind of make this happen so that the number is actually based on my audience so as we move forward you'll understand it's not just clawed out of the box it's a very stupidly simple system that helps Claude or GPT understand my audience and the score is based on audience interest level.
So if we look here, here's what's really interesting. The first talking point is actually the working title. Now, good chance I test different titles, but the idea is I got an unfair advantage and that's what we see here.
GPT Astra and Claude gave my YouTube channel an unfair advantage, 96. And all I do is every time I create a video, I create a new project. So all these AIs have this feature called a project.
And each project is a single video. And I just voice dictate. And then I say about two or three things.
One, just capture talking points based on everything I share in the next. five or 10 minutes. Number two, pay attention to the documentation I've uploaded to the project.
You'll find information on my audience, again, cheers, their wants, their desires, what they're interested in, their struggles, what keeps them up at night and more, how old they are, as well as some of the things that have worked well for my channel and more. So I'm not just using AI out of the box, I'm guiding the AI. based on my channel.
And then you simply have the AI organize the talking points step by step based on what I like to focus on lately as the audience interest level. For me, most importantly, is when I started breaking down this system, like it had to be able to run on my phone. It had to be able to be simple and easy to access.
That meant I could capture an idea like this is my note taker. Right here, my phone, AI, here's the idea, create a project, and then the talking points. And if we break it down to the core essence of what drives success, two things have to happen.
Number one, I think of the circle, right? And in the middle and like on this side is getting viewers to click. that's the ctr so that's the packaging and here's what's cool packaging has never been simpler it's so stupidly simple once you start really focusing on you know what i'm breaking down so like step by step you can focus on and improve and get better at packaging and it will have a massive impact on your channel that's why my channel turned around because the way i made videos in the past literally kept me stuck, kept me struggling, was absolutely impossible.
And there was no way I was going to win until I really started leveraging this. The other half is keeping viewers watching. And if you think about what I'm doing right now, it's like, well, these talking points are based on the viewer interest.
And all I have to do is like create an audio track. So if you're using AI, then what you want is very specific information that you can use to create an AI voiceover and you go from A to Z. If you're doing talking heads, you talk through your video, you organize the bullet points or Astra does that or Claude does that.
You sit in front of a camera or a phone and you capture your thoughts. That's it. It's literally that stupidly simple.
But now it's organized based on the viewer interest. Now, here's the most important thing is my video ideas begin with what's working right now in real time. And this is.
so fascinating to me is because every day I'm also really paying attention to YouTube and the homepage and what's happening on YouTube and what people in my niche, the audience I'm targeting, like let's not get caught up in semantics, please. Let's focus on like, it's not a niche, it's an audience. Instead, let's just focus on I'm paying attention to the channels that are winning and it just registers with me now.
So, Ed Lawrence published a video maybe a month ago and the title was, unfortunately, YouTube something success or getting views or making money with YouTube. So the structure was sort of like my video I published a day and a half ago. Unfortunately, YouTube success is very simple or something to that effect.
And then I noticed that Gabe, I think his channel is called The Process. Now, he published a video, and guess what his title was? Unfortunately, YouTube Success.
Okay, really cool. Now, I don't watch other people's videos too much. I just, I rather focus on what I want to focus on.
I want to tell my story, but I do kind of want to sometimes maybe understand a bit. I watched a about a minute of Gabe's video and he starts off with Einstein and I'm like, that's not the quote I would use.
There's another quote. It's basically everything should be as simple as possible, but no simpler. Kind of like this system I'm using now.
This system boils things down to eliminate all the stuff that doesn't matter. And this is my YouTube religion. And I thought, well, I'm going to make that video now.
I've seen it show up. And then I checked it was an outlier. Other people are leveraging this format.
I did it my way, but here's what's cool is like, I used a thumbnail that was super contrarian, which I've used so many times, and I know it drives viewers to click. It worked six or seven years ago, eight years ago, it worked. about five months ago, it worked three months ago, and it worked yesterday because human beings are wired to respond in certain ways.
And once something works, it'll probably work again in the future when we talk about that packaging. Now, I wanna connect this, and if you're still watching, just let me know, I'm curious. This'll probably be 10 minutes into the video.
But this style of work is really based on outliers, but my way, my ideas, my video, right? And if you're not sure, you can still use AI to help you come up with a structure you're excited about by just having this conversation.
But what's interesting is the algorithm is literally designed to reward videos that are highly related. And I keep getting this message again and again, and it's basically the message is, Popular videos like this one are shown to viewers watching similar videos on YouTube so they get more views.
And it's like, I love that it says so they get more views. And that's the thing that For me, this actually connects to how the browse algorithm works and a study where YouTube wrote about how browse work and they talk about like our system browse looks at the viewer and how many times they've watched a video from the channel.
So when you upload a video and you don't have an audience that leaves you with nothing. destined to fail, to struggle. However, if you think, and you notice the second part of the message, which is how often has the viewer watched a video on this topic?
Okay, now we have something to work with, which directly correlates. to browse algorithm, to the message I see. And now, so what I'm doing is, it's not just that this title and thumbnail is powerful, but there are all these viewers on YouTube that are watching all these videos just like the video I uploaded.
So if you're like, well, the YouTube algorithm won't give me impressions, this is how I've gotten more. It's worked really well for me, is to pay attention and always begin my video with what is the title going to be? What is it going to look like?
And because every day I'm watching YouTube, I'm paying attention to the homepage, and I'm also paying attention to channels and videos that are winning. in my space, I see these patterns and it makes it really powerful. Now, what's even more interesting is not every video that I publish is based on an outlier.
And it doesn't have to be because over time, if we think about that message, about the algorithm. How many times has the viewer watched a video from this channel?
And as that number increases over time, as a channel uploads more videos, they're more likely to see the next video. Because they're starting to create like this viewing, habitual viewing experience where they're less critical of a video that's uploaded from a channel they like and love because they've already predetermined that this is good and I like this channel and they click.
So that starts giving me freedom to cover the topics that I want to cover. Amazing. For example, I published a video really talking about video length.
And the title was like, does video length impact views? And I created it very much for my audience. So remember, I'm thinking about my audience, who they are, what matters to them, and what their interests are.
And I'm like, well, okay, the interest level is more views. And then I thought, well, this is a question that comes up again and again. How long?
How long should my video be? And I combined those. That's the title.
That wasn't an outlier. It did, I think, 18, 19 ,000 views. Powerful.
And then another one of my videos, interesting, it was from my system that I've been developing, publishing stupidly simple videos. And I noticed other videos talking about three hours. Really creators that are leveraging a similar workflow.
Because like... Who wants to create tons of high effort videos when the payoff is so low? But what if we can like reduce that to simplify?
be done with the video in three hours, so I created my version, and that video, I think that's got now over 30 ,000 views. And so oftentimes, it's not just that you're copying a video, it's that you're aware of what I like to call the conversation happening on YouTube, and that's videos, really channels that are winning the audience that responds, the topics they cover, and the overlap.
So when Gabe and Ed publish a video really talking about the same thing that I talk about in my own style. Like, unfortunately, YouTube is very simple.
Well, then it makes it easy for the algorithm to connect and push a video out. And the system ensures, here's the big takeaway, is that when I begin my process now, okay? I'm working to ensure that I've got a video title or just a working title.
And because I'm paying attention to what's happening on YouTube, MySpace and more, I'm far more likely to naturally choose a video title. That's already a winner. You know, I think about like pizza, like in the fact that Gabe and Ed and I created these, we all created different versions of this video, which is unfortunately YouTube success is very simple.
And if you think about pizza, there could be three different pizza restaurants and they're all gonna serve different pizza, even though like it's a pepperoni, mushroom, the crust, how they make the dough. Is it a pan? pizza?
Is it California style? Is it a wood -burning oven? And this is how I think about this.
I get to be creative. And I'm just, I'm pausing here because I want to encourage people to have an open mind. And if you're like, I'm not sure, well...
There are the people that leverage outliers and the other people who don't. And this system is really based on understanding that conversation. And over time, you have more freedom.
But first, like, I need to connect and get new viewers in. And outliers make that. The second part of this is so like this video, let's see here.
This video is currently about 22 minutes long and I'll be done in 10 or 20 minutes. And then immediately I'm gonna edit the video. I might even have GPT edit this video for me.
Just doing the grunt work because it's so stupidly simple now. That saves me a lot of time, right? Instead of me sitting and editing for 40 minutes, it'll be done in 15 minutes, maybe 10 minutes.
Great. And while that's happening, I'll be thinking about crafting and creating a thumbnail, which we know, like we already know this stuff is so important, the packaging. But now, like with the help of AI, it is so simple to create a pretty darn good thumbnail that does the job.
It's not about the aesthetics or the visual quality, how good it looks. It's more about the clarity. Does it stand out?
Is it based on my audience's desires? And if I say, this is the thumbnail template. I want the thumbnail we're gonna make to look sort of like this.
This is the idea. Let's come up with concepts for my thumbnail. Now here's what's interesting.
Guess what's in the documentation that I created. that Clot or GPT already has. It's like, this is how we validate an idea.
And I'll say to the AI, don't give me a thumbnail until you've validated the idea. And here are a couple thoughts. So, like, number one, it has to be super simple and clear.
the viewer has to instantly know it's about my subject matter, YouTube. So I'm always creating and adding a YouTube play button. Why?
It signifies this is, I should care about this. It's about what I, what matters to me. Second, it has to be super simple to understand and read.
Not visually impressive. It needs to say something that matters. Again, think viewer interest.
So some of my thumbnails. Same video, 3x the views. Even crazier is that by leveraging this system, I can allow the AI to hear my video, to go through my video as I'm actually editing.
Or the first minute after the video's been edited, I'll have the AI listen to the first minute and say, give me a concept. based on two, three, or four words that's super short to the point and validate your ideas based on the documentation. And that part right there is what changes the game.
You see, I'm in control. I'm telling the AI, do not give me these ideas that you think are clever but nobody would understand. AI is not a good thought partner.
AI is not going to help you understand a great structure. In fact, AI will mess up the structure. So when we talked about like the talking points, the fact that I tell the AI one.
You are going to organize this on talking points that matter the most to my viewing audience. So when I tell the AI, like, organize the talking points based on what's most important. And then two, maybe if give me a thumbnail concept and validate the idea, then it goes back to the documentation, looks at all the rules and all the ideas, understands it has to be clear.
It has to be based on these things. It can't be any more than four words. All the hard work.
is done I don't have to remember I don't have to move forward and that's why this this process works so very well and then I get a thumbnail and I'm completely done now what I'll do is I'll usually come up with two or three concepts 60 % of the time. Sometimes I'm just like, this is the thumbnail.
It works. I'm done. I'm burnt out.
I don't have much left today. I need to get this out. And it still does well.
For example, when I published the video yesterday, the thumbnail is like, stop uploading. And I know that pattern. Or that video package where the idea is to grow and to succeed combined with stop uploading just works.
So I'm just going to do it again. And sure enough. It worked.
The video is getting recommended. Now I've taken this system and I've improved it and I'm using a tool now called 1 of 10. This video is sponsored by 1 of 10 and I absolutely love it based on my workflow and the ability to spot outlier videos, winners, really quickly.
So what I can do is I can go into the tool. I can pay attention exactly to the videos that have done well for other channels. In fact, I can go in and I can track channels with one of the features, which is what I've wanted to do for a long time, is to be able to see all the channels that are in my space watching videos so I can kind of ascertain, identify my video that I'm moving forward with, and 1 of 10 makes that so very fast and quick.
And the fact that then I can set up my own track channels and add channels that my audience is watching, means that I get to sort in seconds, again, a video title or topic and it gives me ideas. Like I saw this one video and it's like 10 videos got my channel monetized.
Now I know I can make that same video and I've got a different spin on it based on my thoughts. My ideas. It's my idea.
It's my video. But the title is like leveraging a proven format that's already won. So 1 of 10 has made it so stupidly simple.
And you can actually access, test out, and try 1 of 10 yourself. $1. Super program.
Check out the link in the description. And a big thank you to the folks at 1 of 10. I want to mention one critical thing that...
kind of can cause this to fail, and it took me a while to figure this out. So I kind of mentioned how my talking points are scored based on viewer interest. Well, if I forget that, if I don't do that, the AI, especially when I create like, I'm looking at my flow right now, and I'm on like 47, 48, it will often think that the structure of a great story should be like this and it actually works against how YouTube and viewing audiences work.
Because what I want to do is keep as many people watching in the beginning as possible and retain as many people throughout the video, retain retention, and yet... Often the AI will take a video structure and it'll save the very best points for the end. And I recommend that you lead with the very best points because that way the viewer gets something fast, hard, powerful that they can leverage in seconds.
and they're more likely to continue watching, which extends your average view duration. If you remember, we spoke about like the circle and half is packaging, title, topic, thumbnail, really ensuring I have the highest click -through rate, leveraging outliers, and the second part is keeping viewers watching. So a couple parts is number one, The AI will often default to this perfect structure with a big payoff at the end that 20 % of the viewers will experience.
But what happens if you give that big payoff at 17 seconds into the video? People are more likely to have this epiphany, this wow moment and say, wow. I'm hooked.
Like it happens subconsciously and they keep watching. And that's why structuring it in this way is so very powerful. Now here's the other thing about like keeping viewers watching is that I want to go longer, but I also, it's not just about going longer.
I need to hook viewers early and keep them watching longer. And like I want five minutes plus a 7 % CTR. If I can get 7 % of a cold audience, which if you're thinking, I have that.
Well, yes, but it's different as a video gains more and more impressions and views because you're meeting a colder and colder audience and it drops. Maintaining 7 % as a video, you know, starts driving 10 ,000 views. 30 ,000 views, 100 ,000 views is hard.
But what's interesting is if I can get close to that, I can just keep people watching for three minutes. You know, it's just really like maybe the CTR is lower, but the average view duration is a bit higher and it works. Maybe the average view duration is incredible.
People are watching for seven or eight minutes and the CTR is pretty good and that kind of video can do pretty darn good. So the idea is that if I'm not guiding the AI with viewer interest, I'm going to struggle. Now do this.
This video on the screen right now will help you to move forward and really make sense of this YouTube stuff. The next step. Check that video out.
You'll get the tips, tactics, and more than that. You're going to feed a poodle and a kitten. You dig?
I'm Brian G. I'll see you on the next one.
The Hook
The bait, then the rug-pull.
Brian G. Johnson opens by claiming AI gave his channel an unfair advantage, then promises a system so simple anyone can run it from a phone. The rest explains what the system is, why it works with YouTube's recommendation logic, and where the AI will sabotage you if you let it.
Frameworks
Named ideas worth stealing.
05:36model
The Packaging and Retention Circle
Get viewers to click (CTR, packaging)
Keep viewers watching (retention, structure)
Success needs two halves: titles, topics and thumbnails that earn clicks, and a structure that holds attention once the video starts.
Steal forany video planning checklist
04:11model
One project per video
Create a new AI project
Upload audience documentation and what has worked
Voice-dictate ideas for 5 to 10 minutes
Ask for talking points ordered by viewer interest
Each video gets its own AI project seeded with audience documents so the output follows the audience, not generic advice.
Steal forany recurring content workflow
CTA Breakdown
How they asked for the click.
VERBAL ASK
25:44next-video
“This video on the screen right now will help you to move forward. Check that video out.”
A standard end-screen pointer to another video, preceded by a mid-roll sponsor read for 1 of 10 with a $1 trial link in the description.
Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.
Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
A vidIQ study of 4 million 2026 uploads plus YouTube's own built-in guidance, broken down into three drivers behind titles that are actually pulling views right now.
Brian G. Johnson tests the retention myths against his own channel's analytics and finds the real signal behind recommendations: average view duration, not watch percentage.
A creator tracks a jump from his usual 14K-24K views a month to 165K back to one habit: reprogramming YouTube's custom feed to study which titles and thumbnails are breaking out right now.
A YouTube creator whose channel was losing viewers a few months ago breaks down the three-part system, research, an almost-perfect title and thumbnail, and minimal editing, that turned it into a 600,000-view quarter.
00:20
00:48
01:08
01:22
01:47
02:22
02:45
02:58
03:40
03:55
04:15
04:43
05:02
05:22
05:41
06:01
06:40
06:59
07:19
07:38
07:58
08:17
08:37
08:56
09:16
09:55
10:15
10:34
10:54
11:13
11:33
11:52
12:12
12:31
12:51
13:10
13:30
13:49
14:09
14:48
15:07
15:27
15:46
16:06
16:26
16:45
17:05
17:24
18:03
18:23
18:42
19:02
19:21
19:41
20:00
20:20
20:39
21:18
21:38
21:57
22:17
22:36
22:56
23:16
23:55
24:14
24:34
24:53
25:13
25:32
25:52
07:10
12:38
15:58
13:08
07:51
20:51