← The Koerner Office - Business Ideas and Deep Dives with Chris Koerner
The Koerner Office - Business Ideas and Deep Dives with Chris Koerner
He Can't Code but His AI Agents Make Him $5K/Month⏐Ep. #289
00:43:46
Loading summary
Transcript112 lines
- [00:00]
Vadim
In the last 30 days, I made $5,000.
- [00:03]
Chris
How technical is your background? Did you know how to code?
- [00:06]
Vadim
I don't know anything about coding. Nothing about tech? Nothing at all. I still cannot write you a single line of code. And I was able to quit my normal job and I already made five grand in my first month. My life fully has changed. This is easily like a 50, 60k, 70k year job and I have agents doing this for me. I should tell them what to do.
- [00:23]
Chris
Man, my mind is blown right now because I'm thinking of just like all these venture capitalists in Silicon Valley, they're not going to have projects to give money to because what you're doing is something that a similar startup in Silicon Valley would need to raise one to seven million bucks for. You would not be profitable today, period, end of story. If it weren't for these agents. Right?
- [00:42]
Vadim
100%. I could quit along the way, but I'm like, no, I'm put my head down. I'm going to make this work. I'm going to take the bet on myself and be a self motivator to actually be successful.
- [00:51]
Chris
Dude, that's inspiring. I'm Inspired, man. I'm 20 years older than you and like I didn't know half of these tools existed. My gosh, that's a huge find. What other tools are in your tool belt that people have to use? Like yesterday, lots of talk lately about OpenClaw, the new AI agent that can, you know, run your life or whatever. But what if Openclaw is already outdated? Today I'm talking to an 18 year old that just Vibe coded an app. He has no technical experience whatsoever and this app made $5,000 in the last 30 days. He just vibe coded it a couple months ago and it's a pretty complex app. It competes with a company that just raised $50 million. So we talk a little bit about how he built his app. You know, how to vibe code apps. But we also talked a lot at the end about AI agents. What type of agent is better than OpenClaw? A few agentic tools out there that are changing the game. Talks about what prompts we should use to vibe code better apps, what tools we should use. And just super inspiring. I mean, 18 years old, doesn't know how to code. Please enjoy the Vadim. Well, why don't you tell us who you are and what you do.
- [01:53]
Vadim
Hey, I'm Vadim and I'm the founder of Vu Gola, which is a clipping startup for clippers and marketers. And I'm scaling that to be the number one clipping tool by the end of 2026.
- [02:02]
Chris
Okay, so talk to people about clipping like it's about, it's been around for a few years now. Why is clipping important and what does it do?
- [02:11]
Vadim
So for example, all the streamers you see, all the podcasters you see, clippings only become more relevant in the future. That's why I started building my startup primarily for clipping. Because like if you think about clipping, it's purely organic marketing and it's a win win for both your client and the clippers. If like the clippers get paid a dollar per thousand views and then the client gets like a hundred million views, it's a win win. And in the future, I firmly believe though the future is short form content, there only be more reels, more like posts. And clipping is going to be another like branch of this marketing where it's going to completely take over and especially like content rewards. Bop like the things that they're doing and like changing the industry in itself, like that's gonna be just the future. That's why how I got into clipping, clipping essentially just in broken down terms is if you're in a podcast, live stream, whatever you are. So like if you post your own YouTube channel and then you post your own shorts to your own YouTube channel, that's only five channels of reach you have. But if you have clippers, they make up like 50 sub accounts in your name, repurpose all the, like, they clip like a bunch of clips and short content, all 50. Now you have like 50 channels of eyeballs going all to you. So that's basically what clipping is.
- [03:15]
Chris
So I think like the big elephant in the room, the big question mark with clipping is how effective actually is it at bringing short form viewers to long form viewers or listeners. I, I've done quite a bit of clipping. I've had an agency, I've done on wap, I've done it off wap, I've got a bunch of accounts, I've got my verdict. I'm curious what, what you've seen and I know you're biased, I know you're in the business of clipping, but have you seen any sort of like a correlation or a conversion rate, like how many short form views =1 long form view or whatever?
- [03:47]
Vadim
Yeah. So I can't really tell you the exact metrics of the clipping, but all comes down to how you package clipping. Obviously like if you're a streamer you just want more people to know who you are. For example, Clavicular says he has like a thousand plus clippers. So his primarily like all of his clips are going to be like him like quote unquote, the mogging like trend or him like looks, maxing, whatever is. And that's the kind of clips like that the streamer has. But I think clipping, if you want to tailor like you want to get eyeballs from your short form content to your long form is really good for podcasters because if they see your clip and they get value from the clip, like the clip is good quality, they take something out of it on Instagram, on tik tok, on YouTube, on YouTube shorts, they will go to find your channel, subscribe because they're going to want to see more of that content. So clipping like for podcasters it's like pretty good. But like let's say for a streamer and like I just want like you for entertainment, not for value. That's like where the difference is. Because like it's, it's, it all comes down to like are your clips like for entertainment or are they like actually value driven and for conversions. So that's like you have the streamers of the world, then you have the podcasters of the world. And I would argue the podcasters have a higher conversion rate of people going and viewing their podcast than just like a streamer probably would.
- [04:51]
Chris
Yeah. Okay, now I want to get into this app that you built. Tell me, tell me what you built and what inspired you to build it.
- [04:59]
Vadim
Yeah, so it's a clipping, scheduling, captioning to all in one. And why I built this was I start building any startups at all. I want for us to take back like a step back and be like, okay, wasn't me relevant. Next like six to like two years, like six months to two years. Because the one thing I see in the startup community is people build 10 apps and they just like quote unquote, throw crap on the wall and just see whatever sticks. They don't like actually take time to diagnose, research and actually see like, hey, let me build one legit complex software that can just go scale. So I'm like, okay. Club is to be the future of marketing. Everybody's going to need clippers. And the one thing I noticed that other clipping companies don't do is tailoring for agents because the future is agentic systems, pipelines, workflows. So I'm like, okay, what if I just build out like a prototype MVP stage of like just a normal clipping tool, see if people actually want it. And then based. Based on what people say, then I'll start building out the agentic tool, the scheduling, the captioning. So that's how I got into it. And now we have a couple hundred paying users. Now we've generated, I think over like 10,000 plus clips. And like, so far the growth has been pretty steady.
- [06:02]
Chris
So what is your mrr? What is your lifetime revenue? Let's talk numbers. But both, like, how long ago did you found it and what are your numbers today?
- [06:11]
Vadim
Yeah, so current day. So in the last. In the last 30 days, I made $5,000. So the last 30 days I made 5K as like, no coding experience, Nothing at all. I started building in October. I got into Lovable. I don't know, what I was doing was like cursing at the AI. I was telling it, you're not building me the right page. You're not doing this. And like, only then I realized AI, like, if AI gives you something bad, it's a skill issue. Because AI can only act on the mind of context and prompt engineering.
- [06:37]
Chris
Yeah, it's a skill issue with the user, you're saying?
- [06:39]
Vadim
Yeah. So, like, I was like, I was the most newbiest vibe coder you probably, like, ever see. Like, I like asked Lovable to build me something level. Like, did not build what I wanted because obviously I didn't tell it enough details and I would get mad at Lovable. So, like, that's how I started my vibe coding journey. It was on Lovable. Then I moved to cloud code. It was like, it's a pretty complex software. Like taking clips, framing, cropping, sort like it was like complex software. And I like had so much time to quit. But I started like, when I got into cloud code, everything started working out. Like, start picking it up. In the last probably like two months, I actually started marketing it organically on Twitter and it's purely off Twitter so far, which is also like a bad market for clipping. Like, in my opinion, like, I could scale way better on Instagram, YouTube shorts, TikTok, which I started posting now, but. Yeah, but purely on X in the last two months when I actually started marketing it. And like, the, like the free, like the public beta stage, we made $5,000 in the last, like 30 days.
- [07:35]
Chris
Now is that 5,000 from like monthly subscriptions or lifetime subscriptions? What do you charge?
- [07:40]
Vadim
Yeah, So I have three, three paid tiers. So first is like a $9 plan, and then it's a $39 plan. And then it's a $99 plan. So monthly. Yeah, monthly. Monthly recurring revenue. Okay. And, and then we also, I also figured out that the way I can like upsell like because like for if you're having a software business, if ideally your churn rate like a user will go to your software and like if you, if you like it or not, they're going to leave probably within the like the first two to three months, probably one to three months. They're going to cancel the subscription, leave your software. So the way I even maximize my revenue even more is, is when it, for example, when I had like a monthly and yearly plan. And on top of that, let's say they have a monthly, chose a monthly plan. And when they click on the like get monthly plan, a pop up appears and sells them a quarterly plan. So you want to keep upselling them like better, better deals, maximizing how much revenue you can take from your use, from the, like from the user. Because from like one to three months they're going to leave you. And then after like they, they pick their plan, then you take them to a custom checkout page where you upsell them they additional free credits now. So like it's like a one time purchase for like a bunch of more credits. It's like a really good value deal for the user. So that like when I like start, when I did that, I like saw like my revenue kind of like just like, like spike up on one day. I think like one day I made like a like 900 bucks just like in one day after I made like the upselling perks and features now and
- [09:00]
Chris
you're just promoting this on Twitter. You're starting to do short form video, but mostly just Twitter, right?
- [09:06]
Vadim
Yeah, purely Twitter.
- [09:07]
Chris
Okay. Okay. And how technical is your background? Like did you know how to code or is like vibe coding your first coding experience?
- [09:15]
Vadim
I don't, I cannot like, I don't know anything about coding. Nothing about tech. Nothing at all. Like I kind of just learned along the way. I'm like basically like your grandma. I'm not joking. I don't know anything about like coding space. I'm basically your grandma that got into like tech. And that's over time. I just like learned like and now I know like the, the basics of it probably, but I still cannot write you a single line of code.
- [09:35]
Chris
Okay, why did you start with Lovable and would you recommend that people start with it? I know you have a more complex app. Most people are probably doing more simple things. But what's your recommendation?
- [09:44]
Vadim
There honestly looking back now, I would not recommend lovable at all. First of like in my opinion it's too like too expensive. And also the designs and UI gives you. It's like it, it looks like vibe coated slop. So like in the AI space like there's thing called like this can tell if your website is vibe coded or if it's not. It's called AI slob which just means like your website is like garbage and it's obviously made by AI and loadable. Did like a horrible job on the designs. Way too expensive. Like I said like I spent like $500 worth of monthly credits in the span of four days. Maybe that's just a me problem, but I just burned through so much credits. So like if you're, if you're just beginning into vibe coding, I would like. The terminal looks very complex but, but basically all you need to know is like you need a backend server. So like Supabase to store all your data. You need Vercel to like for example you have a domain, you need to host the domain on the server. So you need Vercel for that. And then for your front end that's like your front end code. So you just like open up cloud code in the terminal and you just literally talk with cloud code. It makes the local files for you, you push it onto GitHub. GitHub is literally like a Google Drive for your code and it's basically all you need to build a startup a software. So it's just like those four things and it seems complex at first, but it's not really like it's pretty easy if you get into it. So you can skip all the replicates of the world, the lovables of the world. You can just kind of skip all that and go straight into cloud code.
- [11:03]
Chris
Okay, so Supabase for the database, Vercel for the hosting, GitHub for hosting your code, the terminal app on your computer for actually making the code. What is the AI tool that you're using to actually like vibe code? It is a cloud code.
- [11:20]
Vadim
Yeah, I'm using cloud code and also for example like I'm using cloud code max subscription. So I'm paying 200 bucks a month. And also for you guys that do not know, I just realized this. So when I first got clock cloud code on. If you, if you buy it on the mobile app, it's 250 bucks. I don't know why. But then if you buy it in the Web app it's 200 bucks. Make sure to Buy on the web app. So yeah, I use cloud code.
- [11:41]
Chris
Yeah.
- [11:41]
Vadim
And also whenever my cloud code subscription runs out, I have a second code explained, man. Codex is basically open AI's coding agents. Like how there's Claude, there's Claude code for coding and then there's OpenAI chat GPT and then there's Codex for coding.
- [11:56]
Chris
Yeah, it's 250 on mobile app because they have to give like 20% to Apple.
- [12:00]
Vadim
Oh yeah, yeah, yeah, I forgot about that. Apple takes a chunk out of that.
- [12:04]
Chris
So had, have you been a clipper in the past? Like what got you into this world?
- [12:09]
Vadim
Yeah, so basically when I started doing this like a bunch of people start reaching out to me from like the clipping space. Like well known names like, like do you have clipping experience? Like I, like I never clipped, I was never a clipper. I never did, I never did anything of that before I started building software like at all. I literally just like, like researched. I'm like, okay, I want, like I want to build the actual legit software that is going to be viable. It can actually scale in the future. The future is going to be short form content. Future can be clipping. Everyone's going to need like that. So why not, why not me? Instead of building a simple like walking app or calorie tracking app or like Some simple I iOS apps, what if I just build a legit complex like almost like it's not an enterprise like software but like a legit software web application that like I can just scale and will actually be like viable and like it's going to be like relevant in the future, you know, so that's, that's, that's why we got into it.
- [12:56]
Chris
So you saw clipping as a tidal wave and you said I'm going to hop on this tidal wave, I'll have an easier chance of success.
- [13:02]
Vadim
Right? Yeah.
- [13:03]
Chris
This is here to stay. Now what about Opus, right? Like is that the biggest clipping app in the space?
- [13:09]
Vadim
Yeah.
- [13:10]
Chris
So are you, do you even like look at them or are you just thinking like this market's so big, not everyone's going to use it. There's room for everyone. How do you look at them?
- [13:17]
Vadim
Yeah, so the way I look at it obviously. So I like, I tested out all my competitors first of all, like before I started on my software, I tested all my competitors and started like deep analyzing with myself and AI because like if you don't know your AI can actually go and like do things. For example, I use a tool called Hermes Agent. Essentially it's Open Claw, but on steroids. And there's a thing called Chrome Dev Tools MCP where it can literally go and open tabs for you, take screenshots and analyze. So my Hermes agent, my like you say Jarvis. So my Hermes agent would go analyze, take screenshots, click through buttons. While on my MacBook Pro I'd also be researching and analyzing these competitors. And one thing I realized is first of all, like, they're like, I really. I also, I went through like the customer reviews. But like, one thing is it's too expensive. The clipping process takes too long. Opus Clip also eclipse the same clips at the same time. Like if you post a podcast at Opus Clip, since they're like, since they have a good. It's like a clipping logic. And like, for example, if you clip a minute and five seconds of a video in one clip, you can have that clip and then in clip 2, it can have the same clip. It can be like a minute five and it jumps to a minute 10 and then minute 20. I don't know if this makes sense. Like it reuses the same clip moments in multiple clips. So you're spending credits basically getting the same as that clip. You're not going to repurpose that same clip over time because all these algos look at your content and see if it's the same content. They're not going to boost you at all. So I like, I deep analyze all my competitors seeing like, okay, they're not a gensec. The user is saying, xyz, I can improve in this, this, this way. So let me start building and if I can make my product 10% better than theirs, and heck, like, the market share is so big, why can't, why can I not just take a piece of it as well? You know? So I had that mindset as well, like you said.
- [14:59]
Chris
So how long did it take you to go from, you know, your first prompt in lovable when you first had this idea and you started building it to your first dollar in the bank account?
- [15:08]
Vadim
First dollar, My bank account. I think it was in the beginning of February. I think it was like February 2nd, I got like my first like stripe notification. And by the way, like my product, I did not market it anywhere. I was building in public on Twitter and I saw like, I saw someone bought the creator plan. And mind you, my product probably sucked back then. Like, it probably was horrible, but someone still paid for it and used it, you know, so like I. So yeah, February 2nd, I think that's when I got my first like notification and first sale.
- [15:36]
Chris
And when did you first start prompting it in lovable?
- [15:39]
Vadim
First start prompting in lovable? It was October, November.
- [15:43]
Chris
Okay, so you spent like a few months kind of tinkering with it, learning the hard way. And then first dollar on February 2nd.
- [15:51]
Vadim
Yeah.
- [15:52]
Chris
And then March, your first full month, you did 5,000 in revenue.
- [15:56]
Vadim
Yep, 5,000 revenue.
- [15:57]
Chris
And how much of that was profit?
- [15:59]
Vadim
Profit? I think I have 80. 80, 85. 80% profit margins.
- [16:05]
Chris
That's amazing. Now, are you, Are you using AI credits on the, on the back end when they generate videos? Is that part of your cost or no?
- [16:14]
Vadim
Yeah, so there's a whole system with like the face analyzing, transcript analyzing, looking at the emotion of the speakers, where they're at. Like, everything's basically on API credits. And yeah, I'm basically outsourcing. And like, when I'm pricing my packages, I'm making sure, hey, like, if I package it in this amount of price, this amount of credits, like, Ideally you want 70 to 80% profit margins in the SaaS world.
- [16:34]
Chris
Yeah. And are you using Claude API on the back end, like, for generating these videos?
- [16:40]
Vadim
No, I'm not using Claude, I'm using Gemini API. That's like, that's the most effective because, like, Gemini does a good job at reasoning. And also Gemini also has a vlm, which is basically. So you have LLMs, it's like a language learning model. Then you have vlms, which is a visual learning model. So not only can the AI look at the transcription and analyze which parts of the. Parts of the clips or which part of the video will be good clips, but also look at the video itself and dissect. Okay, who's the speaker? Where are they at? What are they talking about? Oh, it's a podcast. Oh, this is. Oh, this is like a stream. This is like a gaming thing and stuff like that.
- [17:15]
Chris
Okay, now how have you used openclaw throughout the process of building and or marketing this business?
- [17:22]
Vadim
So I started using Opalca, I think a few months. Yeah, I think like two months ago, three months ago, I started using Opal when it came viral. And the way I got my first, like 400, like I before I had like a free plan which, like as a booster boost our company. I was getting burned through credits because people were making multiple free accounts. But the way I got like my first 450plus users, even though like a bunch of them were free, but I got so much feedback from them was through Reddit marketing. So essentially back when I used openclaw, I switched from it, but we'll get back into that later. I openclaw, I made like a marketing agent essentially. It would go on subreddits, it would go on LinkedIn, it would go on X on Twitter, it would see where people are talking about what's the best clipping tool, what's the best AI video tool? And I gave it a bunch of these specific keywords tailored to my SaaS. So hey, look for like users complaining about clipping for users complaining about Opus Clip. Because even opus clip, back in the day they had like a 48 hour outage. And in that moment I told my market, hey, go like, give me a list like bullet, bullet, bullet, their usernames, the post where they're complaining so I can go and shoot them a dm. Hey, I'm building a clipping tool I would love for you to try for free. So that's how I got like my first 450 like free users on my software and got a bunch of feedback on things to improve on, to remove, simplify and stuff like that with Open Clip Cloud.
- [18:33]
Chris
So how many total users do you have today, both free and paid?
- [18:36]
Vadim
Well, I switched from. What's it called? I switched from there's no free plan anymore, so just a paid plan. And right now we have like 267 paying users right now.
- [18:45]
Chris
Okay, so you got, you had a pretty easy time getting unpaid users with your OpenCL agent just by posting or commenting on Reddit. But then your paid users mostly came from X, is that right?
- [18:57]
Vadim
Yeah, because one thing I noticed is like I was giving way too much free value on the free plan, like I was. Because not only is it's not just clipping, it's also scheduling. Like you can like schedule. It's like a whole social media scheduler as well. Unlike these like clipping tools, they focus on clipping and stuff. Like you can clip and schedule natively all in one ecosystem. So like these people will literally make free accounts, clip, schedule, and then they make another free account, another free account, another free account, and I'll get burned through a bunch of credits. And like I'm like, okay, the way, the way I can filter this is let me just make it simple. $9 a month. Instead of the free plan, I'll do a $9 a month plan and that will filter just the people just to test it out and the actual real users that actually want the product and want the software. And mind you, when I had the free plan, my revenue, like I was not seeing any revenue at all. Like maybe just like one sale in like, like a couple weeks or something. And then in March, in March, I think like the end of February, that's when I like I made, okay, I made decision. I'm not gonna have any free plans. I'm gonna make like the $9 plan, like minimum and that's going to filter people. I'm actually going to see if people actually want to use my software. And right when I did the $9 month plan, people started buying it. I'm like oh wow. So like it was a good filter. Like it was a good move in the end, at the end of my day.
- [20:05]
Beehive Founder
Did you know you can build a six figure business with only a 10,000 person email list? I just talked to the founder of Beehive and he told me that there are people with 10 to 15,000 subscribers making more than newsletters that are 10 times that size. Here's how. Well, this one design newsletter with around 15,000 subs doesn't just sell ads. They run two week online cohorts four times a year that generate six figures. And on top of that they're charging companies 20 to $30,000 to get in front of their audience. And that's on one small focus list. This is not about going viral anymore. It's about owning a niche and monetizing it correctly. And Beehive gives you exactly the tools to do just that. Well, like surveys to understand exactly who your readers are or a built in recommendation network to grow automatically paid subscriptions, digital products, sponsorship tools and you keep 100% of the revenue. You don't need a massive audience to run a successful business. You need the right system. So move your email list over today and try out beehive with 30% off for 30 days with code CHRIS30. Because this might be the simplest, highest leverage business that you can build right now. Just go to beehive.com Chris to learn more. That's B E H I I V.com
- [21:23]
Chris
Chris, do you have any numbers as to what percentage of people upgraded from free to pay versus just bounce?
- [21:30]
Vadim
That's the one thing I had a mistake on. But there's thing called post hog which look at your whole entire website and like your analytics and stuff. I did not have proper conversion and data set up from the beginning so I can't really tell you. Like I can probably go my stripe and have my AI analyze all the stripe IDs and the users and see which users upgrade from the starter plan to the creator plan and then from maybe from the creator to the agency plan. I Could probably look at that right now, but I did from the beginning. I didn't have that set up, but I have it set up now. I at least set it up yesterday with post like analytics, but I didn't have proper like data. Like, I didn't have onboarding, I didn't have offboarding, I didn't have any of that. Which, like, looking at this now from like a marketer's perspective, like, what, like what was I doing?
- [22:11]
Chris
You know, I mean, it's like that's the alternative to not starting. Oftentimes people check all those boxes and then they. All that friction causes them to never launch in the first place. So that's usually how I do it too. Like I'm so anxious to get revenue in the door. I don't set up all the necessary tracking things. But in the end it doesn't really make that big of a difference, you know, like a year from now, you're not going to care about that.
- [22:31]
Vadim
Yeah, that's true. It's all about learning the process.
- [22:34]
Chris
Yeah, but would you say that you were surprised by what percentage of people actually upgraded to paid?
- [22:39]
Vadim
Yeah, I was actually really surprised because like, yeah, I'm like, I'm like, because like my competitor opuskill, they have a free plan. I'm like, why not? Just like, why not me Make a free plan. I'll see if like people are gonna go to me and like no one was buying at all. Like, what the heck? So I'm like, okay, let me just make a free plan and let's see if there's a demand if like people actually want the tool, started marketing it more, started talking about it more on Twitter and people started funneling in and then I started getting more emails, more feedback, things that they like, they didn't like what's improve on what to add and stuff like, like that did.
- [23:09]
Chris
Sorry, did you say that Opus did or didn't have a free plan?
- [23:12]
Vadim
They have a free plan. Yeah.
- [23:14]
Chris
Okay, that's interesting. I would have thought that they, that they didn't because clearly your data has showed that it's not a good idea.
- [23:21]
Vadim
Yeah, but thing was, I think the thing with opus, they raised $50 million by the way. And like if you look at their ads, their marketing, it just, they're just burning money just to get big names like, like, like Mr. Beast or I Show Speed just to use the products. If you look at their pages, it's just like well known podcasts around the world. So I'm like, why not? Like I can like literally niche down and go after smaller podcasters, entrepreneurs, affiliate marketers, whoever might be that need clipping. And I, I can niche down and take a look a big like big chunk of the market share as well that way.
- [23:50]
Chris
Have you thought of building an open claw agent that will go to WAP and start DMing clippers, just like spamming them links to your thing, giving them free credits or whatnot?
- [24:00]
Vadim
No, I didn't think of that. But what I did think is so Daniel Bitson, he's the founder of Content Rewards. He was like a clipper back in the day, YouTube shorts guy, pretty successful young lad. And he's like, yeah, Daniel Bitson, V I T T o N. And he's the founder of Content Rewards and I have an idea and I messaged him on Twitter and on messages. Essentially I want to build out a Gentex clipper where like the future of clipping is not going to be like human clippers, it's literally going to be agent clippers clipping for you. Like think about it, they can go take, take, take your video so they can go analyze on WAP what campaign it is, find that video on YouTube, copy the link, like paste it into V, clip it, they download the clips, they add the trending sounds, the B rolls, everything, and then they submit it on their like social media platforms because they can schedule as well. And then the way clipping works, you take whatever you schedule on social media, you take that like Instagram link, TikTok post link and then you just submit it on Content Rewards, you submit it on wop and that's how you start tracking you for your views and giving you like a CPM which they pay you like 10 cents, 30 cents or whatever, like a dollar per thousand views. And like I thought about, hey, why not like content words, build out like an API key or like wa build out like an API key and just build it for agents. So I'm in the talks with them. So like we'll see where the future leads. But that's where I think we're clipping will be like fully agent.
- [25:18]
Chris
Yeah, yeah. I'm thinking of these patterns like 20 years ago, in the early to mid 2000s, you know, the smart thing to do was to build your business Internet first, right? Like if you're starting a business, it doesn't necessarily have to be an Internet business, but it needs to be optimized for the Internet. And then five years later, mobile became everything, right? Facebook launched in 04 and then Instagram a few years later and then the iPhone in 07 and so everyone was saying, like, your business has to be mobile first. And Instagram was really unique because it was the first social media app to be mobile first, whereas Facebook was mobile web first and then they made an app later because there weren't even apps in 04. So it was mobile first for a while. And now like three years ago, it's like your business has to be AI first. Like AI, like it can be a normal business, but you have to lead with AI. And now it feels like today, in the next five years, it needs to be agentic first. Right? And it seems like that's what you're doing, like you're building the framework for your business to be agentic first. Because eventually, like agents are going to be buying these apps and a human will record a podcast and then an agent will just pick it up and take it from there.
- [26:24]
Beehive Founder
Right?
- [26:24]
Chris
Agents will be doing all kinds of tasks. They're going to have a credit card info, they're going to go buy things for us because they know we need it. Like that's what the future looks like and that's how businesses need to build their foundation to be agentic first.
- [26:35]
Vadim
Percent like, and if, if you're a viewer watching this podcast right now, like if you're building a startup or software, whatever it might be, start tailoring and thinking, okay, how can I build it out for agents? Because literally the future is agentic workflows and like enterprise companies. Literally, like your agent is going to go download my agent skill. Like, for example, people are saying a software is going to be dead because like cloud code can build you anything and you don't need to buy any subscription. Well, no, if I have an agent that knows clipping, but then it sucks at scheduling. Instead of building my own scheduling tool, why not like I just outsource and grab an API key and get like this scheduling skill and hook it up to my agent so now my agent can clip and then schedule.
- [27:15]
Chris
I love it. What, what are you building or playing around with on Open Claw today that people can learn from and copy?
- [27:24]
Vadim
Yeah, so I actually switched from OpenClaw. Spoiler alert. So I switched from OpenClaw to Hermes agent, essentially. Hermes. So OpenClaw is like the one thing about OpenCloud I had a big problem with. I don't know if you experienced this as well, but the memory and like the self learning loop on OpenClaw it was just not like I found myself repeating to openclaw time and time again that like, hey, you're forgetting this. You're not doing this. It had contradicting memory and then a thing called Hermes Agent came out. Essentially it's like OpenCloud's Agentic tool where you plug in like Claude code or you plug in open like codecs your subscription to it, it can do things with it. And Hermes agent has like 40 like 40 plus built in skills. They have a built in self memory system. So on Opalclaw you have to build out this memory system by yourself, which I had to do. But Hermes Agent has everything built in natively already within, like within when you install it, it's already built in. So I switched from OpenCloud to Hermes agent and so far it's been incredible. So like I have like a Hermes like CEO agent, like a Brainrot agent. I have a marketing agent, a growth agent and a specific coding agent all tailored for V Gola. So that's like basically the whole entire company and team, you could say behind Vela.
- [28:33]
Chris
So like wow. I mean if you were to like, if you were to hire a human to replace everything that your agent is doing for you, would that be a full time job? Would that be two humans, three part time job? What do you think it would take?
- [28:47]
Vadim
Oh my goodness, I don't even know. I'm spending about like, I don't know how much tokens worth of like Claude code tokens and codex tokens. I'm spending just on coding itself. And then like I have a customer support agent literally setting up post hoc, like post hoc analytics, doing a bunch of metrics. Having cron jobs in real time like this is easily like a 50, 60k, 70k your job just for like one department of Gola, like moderating full time. And I have agents doing this for me all for like 200 bucks a month, 400 bucks a month, like everything for my own behalf. And like I'm the CEO and I tell them what to do and they do it for me.
- [29:22]
Chris
You know man, my mind is blown right now because I'm thinking of just like all these venture capitalists in Silicon Valley, like they're not going to have projects to give money to because what you're doing is something that like a similar startup in Silicon Valley would need to raise 1 to 7 million bucks for, right? And hire that 50 to 100k person over and over, like two marketing people, an admin person. Like you're spending hundreds of dollars a month on agents doing the same thing. Like you would not be profitable today, period, end of story if it weren't for these agents, right? Because five grand a month ain't gonna pay all those bills 100%.
- [29:58]
Vadim
And the thing also like, so if, if you don't understand how cloud code works for of your user. So essentially when you buy the $200 a month plan, they're giving you like $15,000 worth of actual credits and tokens. And if I'm spending that all the time, and then I go to Codex and another Codex $200 plan, that's another like $15,000 worth of coding and like token usage and credits. So essentially these, like, these companies, because not everybody hits their usage limits and weekly limits every single week. Only for me, I'm maxing that thing out every day. Like, I kind of have this like, anxiety almost that if I don't maximize my, my token usage every week, every day I'm losing money, you know. So like, if you're like, if you're on cloud co. If you're on Codex, maximize your output, you're supposed to be like hitting your limits every single week, you know, so because like, you're really spending like 30, 45, $50,000 worth of credits for like 400, 600 bucks a month.
- [30:51]
Chris
Oh my gosh. All right, Hermes agent, that's a huge find. What other tools are in your tool belt that people have to use? Like yesterday.
- [31:00]
Vadim
So, so I literally was researching with my Hermes agent and like, he literally, like talked to me like, hey, you don't have any, like analytics set up nothing. So like, for your software, you want to set up like, not sponsored nothing. Like, it's called posthog. It's actually really good. Essentially, posthog can literally go on your software and it can see where users are clicking, how long they're on a page for what they're going to like, the whole entire user journey. And then you can build out post hog called events, where essentially it's like your whole onboarding flow. Or for me, like the first user they sign up, then it tracks for the first clip, the first download, the first post, the first caption. And Lee tracks every single user's journey, saying, okay, here they cancel their subscription, here they got more credits. Here they got this, that xyz. And it's basically like a whole entire, like analytical tool that you can like fully deep dive and have your agent just running for you. So like, I got post hoc. I'm on a free plan also, by the way. I don't know how, like, I don't know why I would need a paid.
- [31:59]
Chris
How are they making money?
- [32:00]
Vadim
That's what I'm saying. Like, maybe there's like a limit I'm gonna reach or something. I don't know. I literally got a free plan. Put the, Put the code. Like, I just told my agent Hermes, like, put this code, embed it in my software. I grabbed the API, he gave it to me. I'm like, okay, start setting all these analytic tracks. This thing with, like, if you don't know what to do for your software, the beauty of AI, it's supposed to help you and brainstorm with you, you know? So, like, I couldn't come up with all, like, the postdoc events. Like, it literally like went down, like, step by step. Hey, you need to track this, you need to track that. You need to track this. Boom, boom, boom. And like, I was brainstorming with Hermes, and then I'm like, okay, go, go build it out. And it started building it out for me. And the thing with Hermes Agent as well. Like, I don't know if, like, I wish I could, like, I could share my screen right now here, but you can also see in like Telegram, where what it's working on, what tools, what MCP service it's calling, so you can track its whole entire, like, analytic, like its whole journey. And it's coding thing as well in real time. Unlike for openclaw, you would tell it, hey, do like building this thing, and it would just like go quiet and then just spit, Spit something back out to you. So you, you couldn't really know what it was doing, what it was working on. Maybe like it took the prompt the wrong way or something.
- [33:06]
Chris
It's funny because it reminds me of like, my hiring process.
- [33:09]
Vadim
The, the.
- [33:09]
Chris
The key question that I ask all my potential employees is how well do you work in ambiguity? And it sounds like Open Cloud does not work great in ambiguity, right? It needs more handholding. Whereas Hermes, Hermes agent, like, it's just accountable. It's. It has autonomy. It just goes and does things.
- [33:27]
Vadim
And also another thing. So do you know what Paperclip is? Have you heard of the Paperclip? It's like a new AI tool that built out. It's pretty.
- [33:35]
Chris
I can't keep up, Ben.
- [33:37]
Vadim
Twitter is something new every single day. But essentially there's a thing called Paperclip. It's an open source GitHub project. Essentially, it's a whole, like AI orchestration structure for you. So essentially it's like you have like a ce, literally, you literally have a CEO agent. And then it can break down into like your cmo, your cto, and then those agents can have agent Agents, essentially like it's, it's literally a gentic AI company. So you can literally hire a CEO, tell it, hey, hire me this kind of agent, you tell her your goals, you tell her your to do list, you make issues essentially it literally works for you around the clock 24 7. And the thing with Paperclip. So the founder of Paperclip went on a podcast and talked about a maximizer skill. And in my Hermes agent you want to tell your agents if you're using Hermes, OpenCloud, whatever it might be, is you want to build it out. So like hey, if I give you a task, make sure you go and fix it and do it no matter what. Because like if agents they have access to like click on your web, they can like with Chrome dev tools, MCP servers, they can go click on tools, they can take screenshots, they can like if you give it like your login password, make like a dummy account so I can log into Instagram, it can log into TikTok and log into Google. And like I built out a native like custom skill where like hey, if I give you a task, you work on it no matter what. You do not bug me at all until you like figure it out and you literally self improve proof if you run into any work blocks. If you, if you work, if you go into like an unknown arrow page, you yourself figure it out and only like tag me and like message me back if you only need my help like a hundred percent. And the thing with armies it really figures out and clicks on stuff is sending me screenshots proactively now because like I give a task and it runs for me until it literally finishes it. So if you have like agents I would, I would like advise you guys set that up as well that like give them the access they need and just let them run and only ping you when that task is actually done.
- [35:28]
Chris
I call it the Let let them cook prompt. Keep cooking. What other like killer prompts like that would you suggest or, or hacks or tools to get your agent and or LLM to run better or more.
- [35:41]
Vadim
Yeah, so another good thing is like I think like the, the term is called reverse prompting essentially after like when you're setting up your open cloud or for example now my Hermes agent, you want, you want it to ask you questions back. So for example like you obviously talk with it. Like you want to tell your goals and also like be super specific with AI like if you're a vibe coding, no coding Smiths like me like, like, like once I learned this like quote and like embed it into your whole entire workflow. If AI gives you, if AI gives you a wrong prompt, a wrong build, a wrong UI design, whatever. If the AI gives you something that's not your expectation, it's a skill issue on you, it's a you problem. Like, whether you like it or not, the AI can only do so much on what you tell it. Like on context and prompt engineering. Like, think about AI. Like we, like we might have. Like, I had a wrong impression of AI. That AI is so grand. It knows everything about me. Like, it doesn't know everything about me. If I didn't tell IT stuff, it's not going to go on the web and like, search everything. I like, know my personal goals, my ambitions.
- [36:40]
Chris
Even if it did go on the web, it doesn't know your insecurities, your thoughts, your mind, your history, your background, a conversation you had offline. Like, it doesn't know any of that.
- [36:48]
Vadim
Yeah, and that's what I'm saying. So like, once I found this out, like, hey, like, be super specific about the context and prompt engineering. Like, for example, if you look at like one, like this is a good like, side note. Like if, like if you go on a website and you look a designed website and you want to implement it for your startup, literally pull up a notes tab on the sidebar of your computer and start writing like in description. Not your AI, but just like for yourself and a dog. Like, I like this nav bar, I like this header. I like this prompt. And you want to take screenshots of it because your AI can also like look at screenshots. Don't forget that. Like, you want to make a huge note prompt of like, things you, like, take screenshots of the website, give it to your agent and now has so much more context. So rather saying, hey, I like the website of XYZ.com copy it. Like, you have to be so much more specific in how you prompt engineer your, your AI.
- [37:33]
Chris
Yeah. Any other prompts or hacks that you can think of like that?
- [37:38]
Vadim
Ask your agent, hey, what tools and what API keys? What kind of MCP servers? What do you, what do you need so you can be fully autonomous. Because, like, for example, like the, I don't know, like Chrome dev tools, mcp essentially, like, I'll, I'll talk about this again. I did not know anything about it. And like every time I was asking my agent to go take screenshots of XYZ Web or this website, whatever, it couldn't. And then I told, okay, how can we make you like do it? How? What like what skills do you need? What like go look on open source GitHub, repos, go look on Twitter, go do whatever. And then it told me, okay, there's a thing called Chrome do Chrome Dev Tools mcp where I can, where you can allow it on your Mac mini, your MacBook and I can go open tabs for you. I can go do this for you. So tell your agent, like to make it fully autonomous. Tell your agent, hey, what skills do you need? What API keys do you need so I can help you so we can improve our work, like work relationship and our workflow.
- [38:32]
Chris
Yeah, I love that. Any other tools that you would recommend? We've got Hermes agent, we've got Post Hog, we've got Paperclip. Anything else that is just blowing your mind lately?
- [38:41]
Vadim
Honestly, as of the moment, that's about it. Like it's like, it's like I don't want to like over like complicate everything. That's really all you need to like build like a good like startup business in today's world. And literally I have no coding space. I have nothing at all. And I was able to quit like my like normal like high paying like agency marketing job and like build like a decent income for like just the first month of me building like fully marketing my software. And I already made five grand in my first month just like that, you know. So like it's like, it's super like simple. It's simple work but it's not easy work if that makes sense. Like yeah, the time to build a business, it's so like simple. But obviously success, if success was easy, everybody would be successful. Obviously. Like there's like to be entrepreneur, it's still hard and you just have to go over these roadblocks. Like when I was building my software, one little like last thing, like when I was building my software, like I was running to Roblox, like the, the face tracking, the cropping, like everything was like super off because like a clipping software almost completely raised 50 million to build theirs. And who am I to quit my own, you know, and like I could have quit along the way but I'm like no, I'm put my head down, I'm going to make this work. I'm going to take the bet on myself and be a self motivator to actually be successful. And now looking back now I'm doing a legit like startup company under my belt. It's like fully autonomous AI agents, like my life like fully has changed in a span of like four or five months.
- [40:02]
Chris
Dude, it's inspiring. I'm inspired, man. I'm 20 years older than you and I'm like, I didn't know half of these tools existed. You're inspiring a lot of people. Last question I'll ask you. You obviously did a bunch of research in finding this clipping opportunity. What tools or what methodologies, frameworks, whatever can people use to find profitable, uncovered, undiscovered niches?
- [40:26]
Vadim
So essentially you want to like take a step back and like really like for me, for like, for this like part, I didn't use any AI, I didn't use like any, like, I literally just like, I came up with this gig by myself. Like, like, like you want to take a step back and actually look into the marketplace. Like, I don't care if it takes you like a week or maybe like two weeks. Like you actually want to do deep research by yourself. Like, you want to understand the market that you're going to go into. You want to understand what your competitors are doing. You want to understand where the future is headed. In this podcast we're talking about how the future is agentic software, Agentic workflows, building skills and endpoints, webhooks for other agents. And now having that kind of info, you can start thinking, okay, what can I build that every other company or maybe one niche market will need? So like maybe like automate, like automated leads, but like take it, like take that. Okay, automated leads. How can I make automated leads agentic? Okay, I can build out an agent that literally goes and has like a cron job. So it schedule runs every single day and it can go find me leads on my own behalf and then I take another step further and go and message those leads. It can go follow up with those leads. It can go like, you can think about a whole like agentic workflow. Like you, you connect one dog, you think about it. Oh, mind mapping. This is a huge. You want to go and using mind mapping tools and like build one point. Okay, where does this lead to? Goes here. Okay, it goes into here. And you want to mind map your whole entire research, like your winter research thing about whatever niche you want to go into. And that's basically all I'll, I'll say, like it sounds cliche but literally like you wanna, you wanna be the expert in your own field. Like I know everything about clipping. I know where all my competitors doing, where they're lacking what, what they're not doing. And like that's how I'm going to take the edge on all these like multi million dollar companies and competitors on Fukola. So whatever you're building, you want to be the expert in it and really use your own creativity and mind mapping so you fully understand your whole entire market.
- [42:20]
Chris
This is the original Disney like master plan from like 60, 70 years ago. Was this a mind map essentially?
- [42:27]
Vadim
Yeah, essentially that. And that's like, like the thing about this minor map. Like I bet you when Walt Disney was doing this, he probably first started with just like the, maybe just the music or maybe just the tv. Like second.
- [42:37]
Chris
He kept building onto it over time.
- [42:39]
Vadim
Yeah. And he just like started building and strategizing those mind. Okay, I should like this connects here, this connects here. And like you strategize on your own and research on your own behalf, you know.
- [42:49]
Chris
Yeah. Are there any mind mapping tools that you would recommend?
- [42:52]
Vadim
I think Miro. Miro is a good one. That's probably the best one I would use.
- [42:57]
Chris
Okay. Oh, I've heard of this one. Yeah. So this looks familiar. Okay, where can we find you? Where can we find your app or you.
- [43:04]
Vadim
Yeah, so if you guys want to test my app, it's still in beta stage right now and we're like, I'm obviously like making and polishing it out, like working on non stop. You can find out v-AI.com basically v u g o l a I.com and then where you can find me on social media. So I have, I'm mainly on Twitter. I built up my personal brand on Twitter so like 27,000 followers. And then I'm started posting on Instagram so you can find me at Vadim and then under underscore and then stru so S T R I Z H
- [43:33]
Chris
E U S okay, gotcha. We will link to you there. Well, Vadim, this is amazing. Inspiring. Thank you for all your tools. Speaking of mind map, my mind is blown right now. Thanks for coming on, man.
- [43:46]
Vadim
Thank you Chris for having me.