YT YouTube2Text

YouTube2Text → Transcripts → Я 40 дней зарабатывал на ИИ-Милфе

Я 40 дней зарабатывал на ИИ-Милфе — Transcript

by Харчевников · 4,560 words · 712 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00This woman's name is Diana. She is over
0:0250 years old. And together with her
0:03friends, she has gathered hundreds of
0:05thousands of followers on Instagram,
0:07earning over $ 150,000 from her private
0:09communities and advertising. By the way
0:11, here is another friend of Diana, who
0:13was posted to his account by one of the
0:15most popular US comedians, Bert
0:16Kreischer. All these women are part of
0:18the Vintage Lovers project and were
0:20generated by a neural network. In other
0:22words, they don't exist in reality,
0:24unlike the MILFs around me. A month ago
0:26, I reached out to Leo, the creator of
0:28the highly successful AI model club for
0:30the 50+ age group, Vintage Lovers. He
0:33developed over ten AI MILF accounts
0:34under his brand, peaking at nearly half
0:36a million followers on the club’s
0:38main Instagram page and tens to
0:39hundreds of thousands of followers on
0:41individual accounts of his non-existent
0:42MILFs. At the same time, he not only
0:45managed to achieve $ 150,000 in
0:46monetization through selling ads on
0:48accounts, donations, and developing
0:50private communities for fans. Vintage
0:52Lovers also collaborated with Bert
0:54Kreischer, a popular Western comedian
0:55and frequent guest on Joe Rogan's
0:57podcasts, which is a precedent in the
0:58world of AI influencers. I hadn't heard
1:01similar stories in the field of regular
1:03AI models, so I studied Leo's case in
1:04detail, got to know him, and decided to
1:06launch my own experiment. To create an
1:09AI MILF from scratch, launch her social
1:10media, make her popular, and bring her
1:12to a decent income using new mechanics
1:14for managing a model within the 50+
1:15women niche. To repeat Leo's success
1:18and understand how far one can go with
1:20a non-existent mature lady in this
1:21cruel and unfair world. Which media
1:23personalities we will be able to
1:24contact, how many fans we can attract
1:26and how much money we can make, and
1:27which well-known brands will want to
1:29buy advertising on our blog. At that
1:31moment, I didn't even suspect what kind
1:33of results we could reach. Although the
1:36main feature of my experiment isn't
1:37even that. I have already made two
1:39videos about earning money from AI
1:40girls. Many people, after watching the
1:42videos, copied ideas that had already
1:43stopped working. But this time, I
1:45decided to completely change the
1:46strategy. Our task is to create an AI
1:48model that didn't exist before. To
1:50bring her into social media, and then
1:52use advanced AIs like Suno 2 and Kling
1:54to generate viral creatives with the
1:56model in the lead role. In this way, we
1:58will gather reach and followers and
2:00turn our model into a real influencer.
2:02The simplest method of earning in this
2:04sphere is to move fans already gathered
2:06on social media into a private
2:07community and sell them exclusive model
2:09content through messaging. This is also
2:11called chatting. The essence of the
2:13experiment is that content featuring a
2:15woman 50+ cannot be anything but viral.
2:18How many accounts of women 50+ have you
2:20seen who constantly show off their
2:21beauty? I think, none. But this
2:23approach is exactly what makes the
2:25content unique, provoking viewers to
2:26comment and interact with the model.
2:28That is exactly why 50+ models right
2:30now have a much higher potential, both
2:32in terms of virality and monetization.
2:35But we will get to that high virality
2:36and monetization soon. What you are
2:38about to see will shatter your
2:39conventional patterns. So, how do you
2:41create an AI milf look that goes viral
2:43in feeds and brings in a flood of
2:44followers? And this is where the first
2:47pattern break happens, because we won't
2:49be generating AI women in NanoBon Pro,
2:50GPT Image 2, or some underground
2:52Chinese neural networks. No, we will be
2:54using Journey, but first it needs to be
2:56configured. After entering Midjourney,
2:58you need to open the Personalize tab.
3:00Inside, we will see a multitude of
3:02photos generated by the neural network.
3:03We will need to select those that fit
3:05the criteria for generating a milf.
3:06Photos with people in hyper-realism,
3:08without any fantasy or sci-fi elements.
3:10In this way, we will show the neural
3:12network exactly what kind of visuals we
3:13expect from it. And then our actions
3:15become much simpler. While typical AI
3:17creators usually take several photos of
3:19real girls and blend their looks, we
3:20will simply generate them from scratch
3:22using a special prompt. As for the
3:24nuances, the prompt isn't stable, and
3:25to get different appearance variations,
3:27I still need to change details in it
3:28and experiment more. By the way, the
3:30Midjourney subscription itself cost me
3:3235 dollars. This generation method is
3:35generally considered old-fashioned, but
3:37I decided to try it again on LEO's
3:39advice, instead of using familiar tools
3:41like Nano Banan Pro and everything else
3:43I've already listed. Oh, it's done.
3:46Actually, the result is really not bad
3:48at all. One downside is that the neural
3:50network only accepts text. And when you
3:52want to provide some visual as an
3:54example, some reference, you have to
3:55describe it in text somehow, otherwise
3:57there is no way to convey it. As a
4:00result, I settled on four different
4:02looks. Meet: a dark-skinned gray-haired
4:04milf, a hot fifty-year-old brunette
4:06with a European look, a kind woman with
4:08Eastern facial features, and a blonde
4:09with a short haircut and a distinctly
4:11European appearance.
4:12Mina Eduardovna, what do you think, is
4:14this woman already retired, and if so,
4:16where did she work?
4:19This woman is already retired, and she
4:21likely worked as a senior manager at
4:24McDonald's.
4:25Uh-huh. So, these kinds of photos look
4:27like a typical McDonald's senior
4:29manager,
4:29in my opinion. Yes. Lyubov Alexandrovna
4:31, how realistic is it on a scale from
4:33one to ten, or does it give off
4:35something fake?
4:37Quite realistic. An eight out of ten.
4:40An eight out of ten. What is it missing
4:42?
4:43Too much gloss, shine, polish;
4:45something needs to be removed to make
4:47it less artificial.
4:49And where do you think she might work?
4:52Owner of a front-
4:53Alla Igorevna, what would you say about
4:55this appearance? How would you rate her
4:57? Could she become your friend? She’s
4:59too beautiful to be my friend. She’s
5:02kind of...
5:03She’s stunning.
5:07She’s charming. Who are you,
5:10Sash? I’m afraid we won’t be able
5:13to keep working with these AI ladies.
5:15The subscription for the neural network
5:17expired and the payment isn't going
5:18through.
5:19That’s a classic issue, Anna Igorevna
5:21, don't worry, we can issue a "Plati Po
5:23Vsemu Miru" card in 2 minutes, and your
5:24subscription will work again. "Plati Po
5:27Vsemu Miru" restores the ability to pay
5:29for foreign purchases easily and simply
5:31, just like before. More than 200,000
5:33people are already using the service.
5:35It works like this: go to the website
5:36and issue a virtual card directly from
5:38your phone in a couple of minutes. Top
5:39it up with rubles via SBP from any
5:41Russian bank card, instantly and
5:43without fees. You can even top up the
5:45card with a credit card. The grace
5:47period remains. Moreover, "Plati Po
5:49Vsemu Miru" cards. Can also be linked
5:50to Apple and Google Pay. For payments
5:52abroad, you can simply tap your phone
5:54at the terminal, at the checkout, in a
5:56cafe, store, anywhere. And for payments
5:58in China, the cards can also be linked
5:59to Alipay and WeChat Pay. Three cards
6:01for different tasks. A card for
6:03subscriptions, which is why we came
6:04here. Claude, ChatGPT, iCloud, Spotify,
6:07Netflix. A travel card for hotels,
6:09plane tickets, and checkout payments in
6:11190+ countries. And the premium card
6:13has everything the travel card offers.
6:14Plus the most favorable exchange rate,
6:16increased limits, and 24/7 personal
6:18support. There are no top-up fees.
6:21Follow the link in the description,
6:22enter the promo code "ANYA" and get
6:24free 500 rubles. for registration. And
6:26now, Anna Igorevna, it’s time for me
6:27to get back to work. Please don’t
6:29hang up.
6:30As you wish.
6:30Without much thought, I first decided
6:32to settle on a European blonde. In my
6:34opinion, she gives off a vibe of luxury
6:35, she is attractive, and provocative
6:37content with her will look as organic
6:39as possible, but that’s not certain.
6:40Maybe eventually we will switch to a
6:42dark-skinned, gray-haired woman. Pause
6:44the video and write in the comments
6:45without looking ahead. Which milf do
6:47you think will blow up our social media
6:49feeds? In the meantime, please meet and
6:52greet. This milf’s name is Sofia. She
6:54is 57 years old. And starting today, I
6:56am launching her social media. Just
6:58kidding, I’m only launching one
7:00social network, and as a test, I
7:01decided to choose Twitter, also known
7:03as X. Yes, you heard me right. And my
7:05previous AI models 'accounts, the
7:07Vintage Lover accounts authored by LEO,
7:09were blocked by Insta more than once
7:10for unclear reasons regarding platform
7:12rule violations. So I decided not to
7:14make Instagram the main source of
7:15traffic. At least for now. Well,
7:17we’ll manage to get around to
7:19launching Insta too. Twitter has a lot
7:20of advantages. Accounts here don't
7:22require a long warm-up. The platform is
7:24more lenient toward any content, yet it
7:26still has an algorithm for promotion.
7:27Naturally, the organic reach isn't as
7:29powerful as on Insta or TikTok, but at
7:30least it exists, unlike on Telegram,
7:32which is loose on censorship. I don't
7:34know if Twitter will work. But what
7:35else is there to do in the current
7:36reality of blocks and bans? So let's
7:38move on to slowly launching Sophie's
7:40page. What kind of content do you think
7:43we'll create first? What content will
7:45lead us to that long-awaited,
7:47unexpected reach and monetization:
7:50knitting techniques or cooking recipes?
7:53You guessed wrong. The thing is, the
7:54entire 50+ AI model niche is based on,
7:56you could say, the cultural phenomenon
7:58of ageism. It's a biased attitude
8:00toward a person because of their age.
8:01Theoretically, if you're a woman and
8:03you're aging, then you're not beautiful
8:04. By the way, I completely disagree
8:06with that statement. That's why we're
8:08testing formats that sort of provoke
8:10ageists into toxic behavior in the
8:11comments, which means engagement and
8:13promotion of our content in the
8:14algorithms. For example, in her first
8:17videos, Sophie repeats trends with
8:19extravagant speeches from a podium,
8:21flirts with her young boyfriend, plays
8:22golf in a provocative manner, meets
8:24fans at a restaurant, films trendy
8:26lip-sync dances, and simply asks
8:27viewers to rate her look or suggest
8:29what men find attractive in mature
8:30women. And how would you rate her? A
8:38beauty, right? Write it in the comments
8:40. Interestingly, the very first posts
8:42on Sophie's page started getting 3 to 5
8:44, and sometimes even 20,000 views, but
8:45they didn't go any further than that.
8:47In just the first day, I managed to
8:49gather more than 300 followers on
8:50Sophie's Twitter. I also bought a
8:52checkmark for over 5 euros for the
8:53model's account to increase audience
8:55trust. But it didn't really affect
8:57anything. After 10 days of active
8:59posting—videos, photos, and just text
9:01posts—Sophie had a little over 500
9:02followers on her Twitter. No spikes in
9:05reach ever happened, and the posts
9:07themselves started getting not 5-10,000
9:09, but 1 to 2,000 views. But I decided
9:11not to wait for several thousand
9:13followers and posted the link to
9:14Sophie's private community on Twitter
9:15on the fifth day of the experiment. As
9:17I already said, the platform has fairly
9:19light censorship, unlike Insta, so we
9:21have no reason to fear getting blocked.
9:23As a result, by the tenth day after
9:24launching the model, six subscribers
9:26had already joined Sophie's private
9:27group, paying five bucks each for
9:29access to her content. Yes, six fans in
9:31just 5 days after posting the link. 6
9:33times 5, well, 30 bucks. Not bad for a
9:35start.
9:36The nuance is that in 5 days we made
9:38not 30 bucks from these fans, but a
9:40little over 240 dollars. Usually,
9:41creators include the ability to message
9:43the model in the subscription, but this
9:44time I decided to do it differently. A
9:46subscription is needed to view posts in
9:48Sofia’s private feed, but anyone can
9:49message her. And it was precisely
9:51through chatting that we earned the
9:52majority of our fan donations. Actually
9:54, to do that, I had to use a secret
9:56method that makes life much easier and
9:58boosts cash flow. I’ll reveal it very
10:00soon. Here’s a small hint. During 5
10:02days of active chatting with the
10:04audience, I didn't write a single word
10:06myself. Think about it. The experiment
10:08began on July 18th. I posted the link
10:10to Sofia’s private community on that
10:12date, and by July 28th, we had made our
10:14first 240-plus dollars. The next step
10:16was to raise the bar. I shifted to more
10:19experimental and trendy formats, and
10:20also continued to poke at the nerve of
10:21ageism. Lots of photos of Sofia asking
10:23for ratings on her looks, talking in
10:25videos with MILF jokes, even hitting on
10:27Elon Musk. But the reach didn't exactly
10:30grow; it started to slowly roll back,
10:32and instead of the usual 3,000–7,000
10:33views, it dropped to 1,000–2,000. By
10:36August 1st, we reached 344,000 views on
10:39the account, but the number of
10:40followers barely grew. By the way,
10:42revenue from the private feed reached
10:44280 dollars. In short, the rate of
10:45income growth also slowed down. And at
10:48this stage of the experiment, to be
10:49honest, I started to lose heart because
10:51I had originally counted on faster
10:52growth for the model. It happened. We
10:56gathered a decent number of followers,
10:58and everyone who wanted to, clicked
10:59through and bought access to Sofia’s
11:01private feed. And now we need to find
11:05some growth-hack strategy that will
11:07allow us to get not just thousands of
11:09views, not even tens of thousands, but
11:11at least hundreds of thousands. So on
11:15August 1st, I decided to create an
11:17Instagram for Sofia and start warming
11:19it up for the launch. And I also
11:22realized that we needed to create some
11:24more interesting content. Given the
11:27success of conversational formats on
11:28Sofia’s account, I generated more and
11:30more videos with appearance ratings and
11:32various jokes. Yes, we got spikes of
11:3510,000–15,000 views, and this post
11:37with photos even hit 50,000 views. I'm
11:4058 and I have two children. Would you
11:42text me? Would you write back, just be
11:44honest. On August 10th, Sofia’s
11:45account reached 1,300 followers on
11:47Twitter. 22 paid subscribers had
11:49already entered the private feed, about
11:51100 free ones, and the total income was
11:52555 dollars. And yes, I remind you, we
11:54made almost half of this money by the
11:56tenth day of the experiment, and we are
11:58currently on the twenty-third day. It
11:59is obvious that we still haven't found
12:01the growth-hack format that would allow
12:03Sofia to go truly viral. And at that
12:04moment, I finally decided to radically
12:06change my creative direction. Starting
12:08August 10th, I went all out with the
12:10most provocative formats that were
12:11already going viral in the AI girl
12:12market. Here is Sophie flirting with a
12:14customer as a Starbucks barista. Here
12:16she is bringing ice cream to her
12:17boyfriend's car, flirting with him. And
12:18here, her boyfriend is playing with a
12:20remote control for some toy on the
12:21subway. And Sophie is reacting to it.
12:23I wonder what that gadget is? Yes.
12:26One more thing. After two weeks of
12:28working on the model, I got tired of
12:30constantly switching between neural
12:32networks. I struggled, but I found a
12:34solution that solved this problem and
12:36made the process of working on Sophie
12:38much easier. Synthesia is one of the
12:40largest aggregators and tools with a
12:42well-thought-out interface for
12:44generating text, images, video, or
12:46audio. For example, I generate photos
12:48of Sophie and, without going to any
12:50other service, I animate them in SD 2.5
12:52. Yes, the platform now has a new
12:54version of the buzzed-about neural
12:56network, allowing for highly detailed
12:57videos. Moreover, you can pay for the
12:59subscription with any Russian card, and
13:00the service is available without a VPN.
13:02You've all seen the trending videos and
13:04photos generated in neural networks,
13:05like the shots from baseball games.
13:08Syntax also added a "Trends" section to
13:09its website, where you can check out
13:11these viral creatives and make them
13:12yourself in two clicks. Syntax will
13:14automatically pull the necessary
13:16prompts and enable the neural network
13:17you need. Click the link in the
13:19description, enter the promo code, and
13:21get a 15%discount on any plan for new
13:22users. And I’ll get back to
13:24generating content. Yes, you have to
13:26agree, such creatives look a little
13:29strange for our AI model. They would
13:31look much more appropriate on younger
13:33AI girls. In three days, we posted six
13:36highly provocative creatives. Four of
13:38them gathered 3, 4, 5, and 20,000 views
13:40, but the other two really brought the
13:43heat. The videos with the suggestive
13:44gadget in the supermarket and on the
13:46subway allowed us to gather our first
13:47500,000 and even a million views.
13:49Moreover, this result can basically be
13:51turned into a system. Hooray! We
13:53finally dug up that very strategy that
13:55most market participants are missing.
13:58The scheme is simple. Take a 50+ model,
14:00make videos based on references that
14:02obviously fit youth trends popular with
14:04standard young AI girls, and you get
14:05incredible viewer emotions and an
14:07effective traffic funnel. It also
14:10became obvious by this point that it
14:12was time to launch Sophie's Instagram
14:14account. We already have working
14:16creatives that performed great on
14:17Twitter. Plus, we can post a link to
14:19our Instagram profile and drive the
14:21audience there. By August 13th,
14:24Sophie’s Twitter had already hit
14:255,200 followers, 3 million impressions,
14:27and her total private income reached
14:29over—wait, before revealing the
14:31income we reached with Sophie’s reach
14:33, I want to tell you a bit about the
14:35chat. I think it’s time. The thing is
14:38, I haven’t written practically a
14:40single message from Sophie’s account.
14:43Someone else was handling that. But
14:45Sophie was actively communicating with
14:46them. And no, I didn't hire any other
14:49chatters; I decided to use a really
14:52cool life hack. This time, I decided to
14:54enable the built-in neural network for
14:55chatting on the FunW platform. Here’s
14:57how it works. We answer questions like:
14:59how you want the model to address fans,
15:01the model's favorite emojis, what the
15:03model won't do upon a fan's request,
15:04and so on. Then we upload video and
15:06photo content for sale and set a price
15:08tag. In my case, it’s one video for
15:09100 bucks and three photos. One for 30
15:11dollars and two for 10 each. The neural
15:13network analyzes our answers and then
15:15chats with fans, trying to natively
15:16sell them the content. If at first I
15:18didn't believe an AI could
15:20realistically imitate a person, now
15:22I’m starting to doubt. These chats
15:25with the neural network look something
15:27like this. A man writes: "You have an
15:29incredibly beautiful body." Sophie
15:30replies: "Thank you, sunshine.""What
15:32else do you like?""A beautiful smile
15:34and the best chest in the world." And
15:35yes, this guy, according to his words,
15:37works in oil production and loves
15:39playing golf, or here’s this
15:40correspondence. A guy writes: "Do you
15:42like massages?" Sophie answered: "Yes,
15:43I like a good massage." And then added:
15:45especially if you know how to handle
15:47your hands. The guy replied that he
15:50also knows how to work with his, I
15:51apologize, this, well, I’m
15:53uncomfortable even saying it, tongue.
15:56Though, there were moments where the
15:57neural network didn’t handle things
15:58quite so well. Here Sophie asks how
16:02interested the guy is in her. He
16:04replies that his interest faded when he
16:06saw the price of the model's photo.
16:07Sophie answers that she would like to
16:09give him more than just photos and be
16:10closer. To which the guy asks, how so?
16:13Would it make him feel good? The neural
16:14network then replied to the guy that
16:16her brain did a backflip. In short,
16:19maybe I wasn't wrong to outsource this
16:21to an AI; my own psyche might not have
16:23held up. By August 13th, Sophie’s
16:30private channel had already earned
16:321,278 dollars. And yes, that’s
16:33without even counting Sophie’s
16:35Instagram. Though, I’m not yet sure
16:36that things will go as smoothly with
16:37Insta as they did with Twitter. Wait,
16:39by the morning of August 15th,
16:41Instagram had gathered only 300-
16:43something followers. Yes, for 2 days of
16:46posting growth hacks with provocative
16:47toys and funneling traffic from Twitter
16:48, such a result could be considered
16:50good. But the creatives themselves
16:52performed poorly, not getting more than
16:53a couple thousand views. Given that I
16:55took the reference for these videos
16:56from Instagram itself, it is safe to
16:58assume the platform's audience just got
16:59tired of such videos. But by the
17:01evening of August 15th, I saw a
17:02completely different picture. That same
17:04video, which hit 500,000 views on
17:06Twitter, blew up on Instagram to 800k,
17:08bringing in almost 2,000 subscribers.
17:10As a result, it started pulling up
17:11other creatives as well. And by August
17:1317th, I already had over six thousand
17:14subscribers. And that's when I
17:16remembered another experiment I wanted
17:18to conduct as part of this video. Our
17:20mentor Leo, whom I mentioned at the
17:21very beginning, also managed a separate
17:23account with a model named Evelyn, a
17:24milf without a private account. She
17:25earned money only from brand
17:27advertising, making about 12,000
17:28dollars in the half-year of its
17:30existence. Yes, Sophie doesn't fit this
17:32style of development at all. After all,
17:34she already has a private account, but
17:35I decided to test this hypothesis
17:37anyway. I drafted an email and sent it
17:39out to several dozen brands and adult
17:41toy stores. It’s not for nothing that
17:43Sophie is gathering decent traffic with
17:44such tech content. Our goal is to earn
17:46money from advertising. I will share
17:47the results of this experiment a little
17:49later. But for now, let's move on to
17:51the bad news. Yes, I have some of those
17:53too. After the viral creatives, Twitter
17:55started performing much worse. From
17:57August 16th to 22nd, there wasn't a
17:59single post that reached truly high
18:02reach. Most of the content didn't even
18:04gather more than 1,000 to 2,000 views.
18:06In parallel, Instagram continued to
18:08grow due to those very growth-hack
18:10creatives with special gadgets. And by
18:11August 22nd, she already had more than
18:1320,000 subscribers. On Twitter, by the
18:15way, the model had gathered only 6,000
18:17fans. So yes, I was now driving the
18:19audience to the private community only
18:21from Instagram, because Twitter
18:22decisively refused to give any new and
18:24interesting reach. But then I found out
18:26that this problem doesn't just worry me
18:27. Among creators and models, there is a
18:29rumor that Twitter often intentionally
18:31tanks reach for bloggers for 3 to 4
18:33weeks, while constantly highlighting
18:34the paid post promotion feature. In
18:36doing so, the platform essentially
18:37provokes creators into buying ads. We
18:39believe it. Honestly, having two viral
18:41social networks at once would be great.
18:43But let me show you what results I was
18:45able to achieve without a Twitter that
18:47had taken off into space. On August
18:4928th, Sophia captured the attention of
18:5141.2 thousand subscribers and almost 20
18:53million in reach on Instagram. At the
18:55same time, her private account already
18:58has 102 paid subscribers and 347 free
19:00fans. And the total earnings amount to,
19:02drumroll, 3,592 dollars. I remind you
19:06that only 40 days have passed since the
19:07creation of Sophie's Twitter. The
19:09result is impressive. But before I wrap
19:11up our email outreach experiment,
19:13let’s talk about the expenses. 35
19:15dollars for 1 month of Midjourney, 100
19:17dollars for 2 months of HeyGen, 400
19:19euros for a video generation service
19:21via Stable Diffusion, 11 euros for 2
19:22months of a Twitter blue checkmark, 40
19:24bucks for private community
19:26verification, and my free Telegram
19:27project, Inflelu. Yes, I cheated a
19:29little. My free Telegram project,
19:31Inflelu, is a project for creators of
19:32AI influencers and models, and those
19:34who want to successfully build them. My
19:36team and I regularly share useful
19:38guides, updates, and case studies of
19:39successful influencers. We publish
19:42toolkits, life hacks, and techniques
19:43that can lead each of you to similar or
19:45even better results. We also post
19:47useful prompts here. Right now, click
19:49the link in the description and
19:50subscribe to the Inflelu Telegram
19:51channel. Open the pinned guide to help
19:52you dive into the field in a matter of
19:54minutes and study it. After that, check
19:56out the posts in the feed and try to
19:58achieve results. The pinned message
19:59also contains information and terms for
20:00our private community for those who
20:02want to work in pairs and boost their
20:03progress faster. If you already have a
20:05successful model, you can join our free
20:07private networking chat for those who
20:09have achieved results, where there are
20:10just over 40 people. By the way,
20:12that’s how I met Leo, whose case I
20:14revealed at the beginning of the video.
20:16The channel already has nearly 70,000
20:17subscribers. Beyond that, we’ve
20:19created a separate Telegram bot that
20:21contains all our exclusive workflows
20:23and all the popular neural networks we
20:25actively use. Luma, AnimateDiff,
20:27Leonardo, Stable Diffusion 2.5, Qwen,
20:29Kling Motion Control, and others. The
20:31bot includes 100 free tokens, which is
20:32more than enough for your first tests
20:34using our workflows. There are also
20:35affordable paid subscriptions.
20:37Attention: for the sake of this
20:38promotion, we are currently running all
20:39generations at cost price through the
20:40neural networks. Meaning we provide
20:42generations without any markup. Yes.
20:44It’s a special offer purely to bring
20:46more people together, so feel free to
20:48use it while you can. The link to my
20:50Telegram channel is below in the video
20:51description. Welcome. And now, let’s
20:54wrap up the experiment with advertising
20:56on Sophie’s account. Over the week, I
20:58wrote to 74 brands, but I didn't
20:59receive a single follow-up response.
21:01Sometimes I got automated replies that
21:03the offer was passed to a manager, but
21:05more often, I was just ignored. Most
21:06likely, the idea of advertising, even
21:08for extravagant brands on an account
21:09and a model with private content,
21:11wasn’t a good one from the start. And
21:13overall, such a reaction from brands
21:14highlights an important issue in the
21:16industry. The market and the models are
21:17still at the very beginning of their
21:19development. Yes, you can dive right
21:20into it, make a quick buck on private
21:22content, donations, or in perfect
21:23conditions ads, but there is no
21:25guarantee your AI model as a project
21:27will show results in the long run. At
21:28the very least, your girl's Instagram
21:30account can easily get banned. I think
21:32that's largely why advertisers are
21:33still afraid to work seriously with
21:35such influencers. So, it is still too
21:37early to talk about long-term AI
21:39bloggers, but you can earn a lot from
21:41them. The main thing is to do it
21:43consistently, methodically, and by
21:44testing new hypotheses. I have already
21:46tested this three times in videos and
21:48several times in the Influ project. As
21:50for why such a huge number of wonderful
21:52people were sitting around me for this
21:54entire video, I won't say. Write your
21:57opinion in the comments. Sasha
21:58Kharchevnikov was with you. Bye
22:00everyone. This what a

About this transcript

This page contains the full transcript of Я 40 дней зарабатывал на ИИ-Милфе by Харчевников, generated from the public captions YouTube serves with the video. The transcript has 4,560 words across 712 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.