21:45The Next New Thing · Listicle
12 shocking things you can make with Fable
Andrew Warner and Brian Casel tour 12 community builds from Claude Fable 5 — then share the three prompting patterns that let it run deep without hand-holding.
June 10thThe argument in one line.
The real edge in AI-era businesses isn't the product — it's distribution arbitrage and knowing how to use AI tools before your market does.
Read if. Skip if.
READ IF YOU ARE…
- You're evaluating which AI-adjacent business to start and want concrete examples with real revenue numbers and honest critique.
- You've been vibe coding or experimenting with no-code AI builders and want to know which categories have real upside.
- You're a service provider (consultant, agency, freelancer) wondering whether to productize your AI workflow knowledge.
- You're curious whether organic social, SEO, or x/Twitter is the right distribution channel for an AI tool product.
SKIP IF…
- You want a step-by-step how-to for building any of these businesses — this is analysis and debate, not a tutorial.
- You're looking for businesses that can scale to $10M+ ARR without caveats — most of the examples here have a ceiling.
The full version, fast.
Eight AI-powered business models are put under the microscope by a founder and an investor: a Claude Code-built power dialer, a vibe-coded app builder competing with Lovable, a fractional chief AI officer service, a viral-shorts tool at $600K/month, a niche fitness app at $20K/month, an AI video explainer platform, a niche SEO directory, and an ad-monetized utility web app. The recurring verdict: the software barely matters — distribution is everything. The one model both hosts agree is the easiest and highest-ceiling right now is the fractional AI officer, because the market is massive, the sales motion is simple (LinkedIn outreach), and it can convert into recurring software revenue.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 00:47
01 · Cold open — montage of founders
Fast-cut montage of eight founders quoting their revenue numbers: $20K/month, $1M ARR, $4.5M, $600K/month. Sets the promise: these are real people who did it.
00:47 – 05:30
02 · Playbook 1 — AI power dialer (Clarvo, $1M ARR)
Founder used Claude to generate 200-300 product ideas within the constraint of 'improve call pickup rates for local service businesses.' Most ideas were garbage; one became Clarvo. Key insight: AI ideation works when you give it a container (defined problem + metric), not an open-ended 'come up with business ideas' prompt.
05:30 – 10:20
03 · Playbook 2 — Vibe-coded app builder (Shipper, $307K ARR)
Two non-technical founders built Shipper (like Lovable but adds mobile apps and Chrome extensions) at $25.6K MRR. Distribution secret: one-liner tweets on X that get 1M+ views by latching onto new model releases ('Opus 4.8 is so good, it's scary'). Hiten's take: differentiation is temporary; the real moat is their distribution on X.
10:20 – 15:00
04 · Playbook 3 — Fractional chief AI officer ($4.5M claimed ARR)
Consultant identifies AI opportunities inside companies, implements them, then charges a recurring maintenance fee. Hiten flags the ARR claim — services revenue that churns isn't truly recurring. But both hosts agree the underlying model (find problem, implement AI, stay on retainer) is strong, especially with the transition toward installing a private OpenClaw-style bot in clients' Slack.
15:00 – 15:30
05 · Sponsor — Zapier MCP
Andrew pitches Zapier MCP as a single toolbox that works across all AI clients (Claude, Hermes, etc.) with clear access restrictions.
15:30 – 18:30
06 · Playbook 4 — AI viral shorts tool (Revit, $600K/month MRR)
Revit makes AI-generated viral short videos and grew by catching every new AI video model wave. SEO is primary acquisition. Peter Yang (Tibo) interview clip. Hiten's flag: AI video generation is expensive — costs could be $400-500K on $600K revenue, leaving thin margin. His framework: only enter a business if you have a distribution advantage.
18:30 – 21:35
07 · Playbook 5 — Niche fitness app (CutCoach, $20K/month)
Solo founder built CutCoach (weight-cut assistant for combat sports athletes) in one month with Cursor and ChatGPT. Beta-tested with his wrestling club. Organic TikTok/Instagram posts at 200-500 views drove 10-15 downloads/day — high-intent signal. Hiten's ceiling estimate: $100-200K/month max because the addressable market is narrow.
21:35 – 25:50
08 · Playbook 6 — AI video explainer platform ($100K/year)
Non-developer built an AI explainer video platform using Base44 in two months ($100K ARR, 5,000+ users). Positioned as the 'all-in-one' alternative to stitching together multiple AI video tools. Distribution: organic LinkedIn only. Hiten: the real arbitrage is the knowledge gap — people who know how to get AI video to look good vs. those who don't.
25:50 – 30:06
09 · Playbook 7 — Niche SEO directory (Luxury Restroom Trailers, $737/day)
Founder vibe-coded a directory for luxury restroom trailer rentals using Claude Code and an open-source Crawler AI repo. Topical authority across thousands of geo-specific pages = unfair SEO advantage. Distribution: pure SEO. How to find other niches: Ahrefs/Semrush keyword research for underserved high-intent terms, then build a directory around them.
30:06 – 32:20
10 · Playbook 8 — Ad-monetized utility web apps
Vibe-code simple utility tools (mortgage calculators, typing tests, business name generators) and monetize with Google AdSense. At 1M monthly visitors, expect $3-6K/month. Build fast with Horizons AI builder using plain English. Hard part: still need SEO or promotion to get Google to rank you. Low floor, low ceiling, easy to build.
32:20 – 33:27
11 · Verdict — which one would you start?
Both hosts land on Playbook 3 (fractional chief AI officer) as the easiest and most defensible business right now: massive market, simple sales motion (LinkedIn outreach), natural path to software. The AI arbitrage frame ties all models together.
Lines worth screenshotting.
- The product category barely matters — the founders winning are the ones who solved distribution first.
- When AI tools make every market massive, the edge goes to whoever can reach customers, not whoever builds the best product.
- Mining AI for 200 ideas and filtering down to one is a legitimate product discovery process — it's not ideation, it's optimization within a container you already defined.
- Competitors like Lovable and Bolt don't hurt smaller players — they create the market and every dollar of their marketing lifts the whole category.
- ARR means the same customer keeps paying. Revenue that looks like ARR but has high churn is annualized revenue, not recurring revenue.
- A fractional chief AI officer is the easiest AI business to start today: LinkedIn outreach, lukewarm leads, no product to build, and it converts into recurring maintenance fees.
- A niche mobile app at $20K/month can be built in a month with Cursor — and 10-15 downloads per day from 200-500 view posts is a real signal, not a vanity metric.
- AI video generation has a cost problem: a $600K/month business can have $400-500K in model costs, shrinking real margin to a fraction.
- SEO directories are back — vibe coding makes them cheap to build, and topical authority across thousands of pages is an unfair SEO advantage.
- Ad-monetized utility web apps (calculators, typing tests, generators) are simple and proven: a million monthly visitors yields $3-6K/month, but the hard part is still getting Google to rank you.
- X/Twitter is the new website: skill sets that used to require a developer now require someone who knows how to prompt — and you can monetize that knowledge gap.
- The AI arbitrage insight: learn something valuable about AI that companies need, charge them to implement it, then stay on retainer to maintain it.
Eight AI businesses, one distribution rule
WHAT TO LEARN
Every model in this video lives or dies on one thing — not the tech, not the AI, but whether the founders found a channel that actually reaches buyers.
- Give AI a tightly defined container when brainstorming solutions: specify the problem, the customer, and the metric to optimize — then let it generate volume.
- Competing AI products that are bigger than you are an asset in early markets: their marketing spend validates the category and warms up your future customers.
- When evaluating any AI business, ask what the costs are, not just the revenue — AI video generation at $600K/month can carry $400-500K in model costs.
- Services revenue is not ARR unless the same customers keep paying. Annualizing churn-heavy revenue is a presentation choice, not a business metric.
- The fractional AI officer model is the lowest-friction AI business available right now: LinkedIn outreach closes your first customer, and implementation converts to recurring maintenance.
- A niche app with a 200-view post driving 10-15 daily downloads is a stronger signal than a viral post — high-intent traffic beats high reach in narrow markets.
- Niche SEO directories get topical authority for free: hundreds of geo-specific pages around one narrow topic earns Google trust no single-page competitor can match.
- Ad-monetized utility apps are easy to build with AI but hard to grow — the soft part (prompting the builder) is solved; the hard part (ranking on Google) is unchanged.
- The AI knowledge gap between companies that have implemented AI and companies that haven't is currently the best arbitrage available to anyone with hands-on AI skills.
Terms worth knowing.
- ARR (Annual Recurring Revenue)
- Revenue that is contractually guaranteed to repeat each year from the same customers. Often misused to mean any revenue annualized — the 'recurring' part requires actual retention, not just a high top-of-funnel replacing churned customers.
- Vibe coding
- Building software by prompting AI tools in plain English rather than writing code directly. Associated with tools like Lovable, Bolt, Base44, and Replit.
- Base44
- A no-code AI app builder (similar to Lovable/Bolt) that one founder in the video used to build a $100K/year AI explainer video platform in two months.
- Power dialer
- A sales tool that automates outbound calling to maximize calls per hour and pickup rates, primarily used in local service businesses like HVAC and roofing.
- Fractional chief AI officer
- A consultant who embeds in a company on a part-time or project basis to identify AI opportunities, implement workflows, and maintain them on a recurring fee — analogous to a fractional CFO but for AI transformation.
- AI arbitrage
- The gap between people who know how to use AI tools to produce valuable outputs (video, code, automation) and people who don't — currently exploitable as a service or product business.
- Topical authority
- An SEO concept where a site that covers every angle of a narrow topic (e.g., 1,000 pages on luxury restroom trailers) earns disproportionate ranking power from Google compared to a site with one or two pages on the same topic.
- Forward deploy engineer
- A model used by companies like Palantir and OpenAI where a technical employee is embedded at a client site to implement and maintain the vendor's software — what the fractional AI officer role resembles at smaller scale.
Things they pointed at.
00:22productClarvo (AI power dialer)
15:30productRevit (AI viral shorts)
18:31productCutCoach
21:35toolBase44
25:58toolCrawler AI (open-source repo)
30:06toolHorizons AI builder
01:03productTypeAhead
Lines you could clip.
01:42
“After mining Claude for 200, 300 ideas, a couple of them were actually pretty good.”
06:30
“Every dollar they spend on marketing is valuable if you're in that category to you as well because these are all new categories.”
12:10
“I just wanna be specific here. There's a mental model where it's like things that took humans and cost hundreds to thousands of dollars are now able to be done for literally a fraction of the cost if you know how to use the tools.”
28:20
“Learn something that's valuable for companies, not individuals, but for companies around AI, and go try to go sell it to them.”
16:00
“I'd only get in any of these businesses, frankly, Andrew, if I had a distribution advantage.”
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.
metaphoranalogystory
00:00How did you make money with AI? I built a mobile app that makes me $20,000
00:04per month. So we just hit a million dollars in ARR. From zero to 50 k monthly.
00:084 and a half million in annual recurring. You're about to meet eight people who did it who will show you how they did it. We've built an AI coding app without any coding experience.
00:17By the way, I don't know how to code. Right? This is my first experience vibe coding.
00:20The entire infrastructure is managed very easily when you do vibe coding. We'll watch their videos and talk about how you can do it too. Presented by Zapier,
00:29the AI automation company.
00:30Heaton, I want you to take a look at this founder right here. Listen. So we just hit a million dollars in ARR with our SaaS product, which we use Cloud Code to build.
00:37And it's the SaaS. It's called Clarvo. It is essentially an AI enabled power dialer.
00:42And just to unpack those words, what this does is it allows us to make more calls per unit time and to have more of those calls picked up on the back end. Then that works really well and is very powerful if you're in an industry that is traditionally pretty call based. So I wonder what you think of this business.
00:59I'm gonna play a couple of clips here, and then let's talk about him. So how do you actually use Claude here? Well, I should note that we didn't actually know how to solve this problem when we started.
01:06Uh, we actually had Claude walk us through every possible way that it knew of to improve pickup rates and increase the total number of calls we could make per unit ton. And most of the ideas were absolutely trash. But after mining Claude for 200, 300 ideas, a couple of them were actually pretty good.
01:22And so the process, if you're interested, is we literally said, hey, we're building insert product here. You know, it is in our case an AI power dialer for local service businesses like HVAC, plumbing, roofing, etcetera. Our core metric to optimize is call pickup rate, which is defined as the percentage of dialed numbers that result in a
01:41Okay. He specifically told Claude, go fire up a bunch of agents, and with these constraints, come up with ideas. He said most of the ideas are absolute garbage, and then he hit on this one idea that became Clarabo.
01:51Do you think that's a good in your experience, having invested in companies, launched companies, is that a good model to do, to to launch? I think the part that's really valuable
02:00that he described was the fact that there were 200 ideas that he had Claude sort of go after.
02:10Right. And then I'm assuming he or with Claude's help, dug into the ideas and found Yeah. The good ones.
02:17So that part is a process I use for almost anything, is because with AI, we can generate more concepts or ideas or just things. So generate a lot of them, and then I would just use the AI to help find the right one based on my judgment and criteria that I'm feeding to the AI.
02:35Did you do that with Type Ahead, the business you recently launched? Yeah. A lot of the a lot of the way that we thought about this product and the copy, and the reason it lands when people go there, is because it's all fully baked on research and options and different things we tested
02:50with the AI to find the exact right words. Like, this whole idea That's copy. But did you actually come up with an idea because Claude Code gave you 20,000 ideas and you picked the one that made the most sense?
03:01No. But this is different. So Andrew, like, we have to go a layer deeper.
03:05Didn't say I came up with a business idea. That's not what he said.
03:09He said I came up with the way to optimize this product idea I already had. That's what he said.
03:14Got it. That's very different. I'm not saying you should have Claude come up with 200 business ideas and you pick two or three.
03:20That's fine, but that that's nuance. Here, think of it this way. He gave AI the container, and within that container, the AI decided
03:29to come up with all these ideas and he picked one. That's my interpretation of what I heard. It wasn't You're 100% right.
03:33I told you that nuance. That makes sense. Okay.
03:36And he basically said, look, here's the problem that my customers have. Here's the number that we're trying to hit for them. Help me come up with a solution for it.
03:42Alright. So AI is good at that kind of ideation. Let's go to the next person.
03:46I wanna know what you think of this one. Here is I interviewed this person.
03:50He's fantastic. Listen to what he does. Buildshipper,
03:53which is an AI app builder. Essentially, it's
03:57Essentially, it's kinda like Lovable or Replit Yeah. But with some more, and we'll talk about what they do.
04:03Here, actually, how much revenue this thing is making. So we're currently at 25.6
04:08k MRR as you can see here. Net volume is 65 k. That's last month plus what we did this month.
04:14ARR is 307 k. So the majority of our MRR is here, but we've migrated Stripe accounts just
04:21I love that he migrated Stripe accounts. Really, as a founder, you just wanna show people, look, I have these both bank accounts. They really are all all connect up.
04:31He's saying that he's different from Lovable because of this. So most of these other apps, you can only build websites or web apps at this moment. With Shipper, you can build websites, web apps, mobile apps, from extensions, and all sorts of bot.
04:44Okay. And? And we're both nontechnical.
04:47We never really like to code. So instead, we've decided to learn marketing. Our first real success was a luxury goods of any
04:54Okay. And then finally, I have seen their ads. Here's where they they promote.
04:58Not ads. I have seen them. We got our first user through Product Hunt, Reddit, SEO, and then x Twitter.
05:04Have you seen them on x by the way? No.
05:07Haven't. They are so good at getting my attention. They will get over a million views on their tweets and it's something like Claude four point eight is so good, it's scary, or Opus four point eight is so good, it's scary.
05:20And then they will link over to their to their other tweet, which says that you can use Shipper and Opus to build the site. Here's the thing that I wonder about this. They are very similar to these bigger companies.
05:32What do you think about non nondevelopers competing with Lovable and all these other businesses?
05:37Is is that an option?
05:40Well, these products have made the market. So Lovable and Bolt and Emergent have essentially made the market by essentially marketing to the market.
05:54So every dollar they spend on marketing is, like, you know, valuable if you're in that category to you as well because these are all new categories.
06:03K. So this one to me like is in this category where I have this concept where x is like the new website.
06:11And when I say that, what I mean is we had a lot of website creators back in the day. They were all different shapes and sizes. You had your neighbor who could h do use HTML and Dreamweaver to be able to build a site for you who might be in your thirties or forties at that time, this is when I this is gonna date us.
06:34This is when we were younger. We would be that person. There were people like us that knew all this stuff about how to build a website, and their neighbor who's a dentist needed a website.
06:43So Okay. In some ways, the way I view and and I and that is I don't even mean the website creation or any of that. It it's really the idea that there are people that have skill sets because of the knowledge they have, and they're manifesting those into helping other people.
07:00So these folks being non engineers are essentially doing some version of the same thing because the market is just so massive, at least that's what people think. And with Lovable's revenue, Bolt's revenue, emergence growth rate Okay.
07:14Like, these things are basically saying that there is just a massive market out there. So that's all to tell you that the reason this is working is because they are good at getting the distribution. I don't know what their churn looks like.
07:27I don't care about the differentiation they said because it's temporary.
07:31Right.
07:32So that's not gonna last. And with a few tricks, you can make all these other tools make Chrome extensions too. So is that really the thing?
07:41So the hidden secret here is the fact that somehow they got to 16 k followers on the account, but it's likely this was an old account. Maybe.
07:50I don't know.
07:51This is the account for their business. Individually, they have good they have stronger accounts, but also they're very good at getting attention. They're very good at these one liners that tap into what's going on on x.
08:02Okay. So what you're saying is, Andrew, it doesn't matter if they're these bigger players. In fact, it's helpful that they're these bigger players that are doing essentially the same thing.
08:09The differentiation in the software doesn't matter. They have a really good chance because they figured out this one type of marketing to this type of person on x on Reddit. And I don't know how long that'll last, Andrew.
08:21And even that line you said
08:23is like think of it this way. If they say, I think they're more careful from what you're saying, but if they say these crazy lines and that gets attention, cool.
08:33Yeah. If their product can do crazy things or say it does crazy things and they can't do it, they're gonna have a high churn. So it's likely this actually has much higher churn, you know, than I used it.
08:45No. The Vision one does well this because they built it on on Claude's software, essentially, on on those models. Alright.
08:52Let's go to the next one.
08:54Oh, this guy I interviewed. Listen to how much money this dude is making. I think people can copy this.
08:58Tell me what you think of model. How much money are you making? You're sitting at about 4 and a half million in annual recurring.
09:04I Here's what he does. Fractional chief AI officer model,
09:07and
09:08Let me just emphasize that. Fractional chief financial chat Fractional chief AI officer, essentially, is what he is.
09:15He's the guy who comes into your company and says to you, the way you're doing things is old. There's a new AI way. Let me help integrate it.
09:23That's not ARR, Andrew. That can't be ARR.
09:27You know what he ended doing? Check? Yeah.
09:31You know what? I think you're you're you're right. They always say it's ARR, and it's not.
09:36Let's let's hear what he said in the video. Good call. How much money are you making?
09:39We're sitting at about 4 and a half million in annual recurring. But but he's charging subscriptions now, but you still don't believe that that's really ARR.
09:49Why?
09:50If it's 4 and a half million dollars of ARR, okay, there's two two things. Yeah. I can I can say I believe that?
09:57Fine. Is it recurring? What is the churn?
10:01Like, those are the kind of questions. It's not usually, this ARR is not the same as, like, full 100% software ARR. Right.
10:09So the second piece of it is he's actually just augmenting human with AI and doing a services business. You and I both know it's easier to get a services business to higher numbers than a software business these days, with caveats.
10:24So my my criticism here, if there is any, is this is not a software business. If it was vibe coded, he's teaching people how how to use AI and all that. That's what everyone's trying to do, and everyone wants that because everyone's trying to do it.
10:37So in some ways, that last idea was lovable, bold, drafting on the marketing and the market those products have created, those businesses have created.
10:48This thing is drafting on the fact that everyone wants to implement AI. Right. So same thing, just very big market.
10:57They're taking their chunk of it and he's probably figured out some distribution on it, but this is a services business. What do you think of the fact that it's a services business though? I don't think that's bad.
11:06I'm just saying we should call it what it is. I don't know if that's ARR.
11:09I just don't know. I think you know what? I think you're right.
11:11There are a lot of people who use annual revenue and ARR in the same way because they assume that if they earn this money last expect to earn it this year, they're going to next year because if they get customers who leave, they're Recurring
11:25means the same customer keeps paying. Right. You get churn if you get churn and not the same customer's paying you, but you have a high top of funnel or something like that, great.
11:36But if they pay you for three months, is it ARR is annualized recurring revenue. The recurring part is the part where I usually have a bigger gripe than you annualize because annualize is by nature times 12 of whatever they're looking at.
11:51But it only is times 12 if it actually recurs.
11:54My argument here is I don't know. And Okay. Fair.
11:57Let tell me what you think. Alright. So fair about that.
12:00Tell me what you think about his business. Here's essentially what he does. Like I said, he sees how your business operates.
12:05He comes in and he becomes your chief AI officer. He implements software. He comes back in and makes sure that the software is hitting your goal, so every month you're paying him to do that.
12:15His next thing is he's installing a version of OpenClaw that he created that is built on the NVIDIA's model of OpenClaw that you will feel more safe in, and that way, he's shifting into software, but he's not now.
12:29What do you think of this as a business, as a way of making money with what you know about AI?
12:35I think it's great, especially if he's gonna transition to to something recurring. I think some of the ways he's thinking about it, I can poke holes in, like the Nvidia more secure, less secure, like, I could argue that five ways No.
12:47To It's just like, it's a narrative.
12:51The narrative is Nvidia's version of OpenCLOS more secure. I'm not saying whether it is or isn't, but the point is even coming at it from I'm doing my version of OpenClaw off of this thing, that doesn't matter.
13:03It's just software running a bot for you in Slack.
13:09How the software runs, nobody really cares about. They don't really care that it's OpenClaw except that OpenClaw is marketing today, and Nvidia's version of it is more secure, which it probably is.
13:19But does any of that matter to the person that you're selling to? Maybe. But it doesn't really matter how it's running.
13:25So I guess it's really awesome that he's found a a thing to go sell to companies that can make money that's like helping them with AI. Moving to this model makes perfect sense if you if you think about what he said, which is I go in, I help find opportunities for AI, I help them realize those opportunities, then I have a recurring fee to help them maintain what I created for them.
13:48Yes. That should ideally be sitting in some software. It's essentially a mini version of what OpenAI and Claude and even Palantir are doing with their forward deploy engineers.
13:58And that's the concept here. So he's a forward deploy AI transformation officer. Cool.
14:04Like, our volunteer does that defense. You the business.
14:08Claude tries to do it for every area they can. Even OpenAI is trying to do it. All the consulting companies are trying to do it.
14:12So again, massive market makes sense.
14:16There's money to be had if you're doing new stuff like this. Let's go to I like your skepticism.
14:22I honestly thought the two of us were gonna gawk at these businesses and go, oh my. This is great. This is this is what people can copy.
14:27I didn't think you'd be this skeptical about them. I like it. Keep it up because I'm the person who's, like, like, cheerleader for entrepreneurs.
14:33Oh, by the way, speaking of cheerleader for entrepreneurs, my sponsor is Zapier Zapier MCP with one simple thing, collection toolbox. I get to bring Heaton everything into all the AI tools that I use, and one day I might be on Claude, the chat app. The next day I might be on Hermes agent, which I freaking love lately.
14:52The next day, I might be on I don't know what else is coming out, and what I like is I have one collection of tools that I can give them with clear restrictions so that I don't have my Hermes agent sending messages to Heat and Azmi by accident. Go check them out at zapier.com/mcp. Next, Peter Yang had this interview.
15:10Do you know this guy with Tebow? You've seen him on Twitter. Yeah.
15:14Yeah. Yeah. Here.
15:15Listen to how much he's making with I'm just gonna highlight one of his Doing a little bit more than 600,000 per month right now, which is And that's insane. With It is insane.
15:26600,000 a month. That's just one business. That's Revit.
15:29Here's what Revit does. Revit,
15:31which is a tool to make viral shorts on social media. And it was right at the beginning of AI videos. And so it took off multiple times.
15:41Like, at every new model, increasing the capabilities of AI videos, it just got way better. And so it's, uh, doing a little bit more than 600,000 per month right now, which is, uh, insane with the team.
15:55This is great. Here's how he's getting his customers. SEO is the number one acquisition channel for Revit.
16:02So it's it's total.
16:04Search engine optimization. He's creating AI videos, and let's talk about how he's using AI to fast forward.
16:10Honestly, I have a have a very basic setup. So
16:14I'm using Cursor. I'm constantly experimenting with the best models.
16:19Like, right now, my my go to models are GPT 5.4 for, like, all the go.
16:29Like, the the complex stuff.
16:31Let me pause right there. What do you think so far of what you're seeing here? I wonder what his costs are.
16:37Oh.
16:38Because you're you're saying that because he's relying so heavily on AI to create these AI videos, the cost might be high. I just wanna know what the costs That's the that's the piece that I would dig into. I mean, it's all a fair business and all that and, you know, like, you know, to to hit on something you said earlier, like, I'm I'm not a cheerleader for founders.
16:59I'm a coach. If you're the cheerleader, I'm the coach. What they need to know is, like, the good, bad, ugly of what's happening here.
17:05So the good here is that he's found something that AI has a cost advantage to humans doing it, which is essentially making these kind of videos. Right?
17:15Mhmm. Okay. Cool.
17:17But if he's making 600 k, I wanna know what his costs are because the cost of making a video with AI is very expensive. So I doubt the cost is 500 k.
17:28I could be wrong. But the cost is probably, like, $2.03, 400 k in that Okay.
17:33In that thing. And if not if I'm more on the aggressive side, if they're using the better models, cost could be closer to, like, $4,500 than it is 2 or 300 k just for producing those videos.
17:46I don't know the answer, but that's yeah. If the costs were expensive today, would you still get in the business? Would I'd only get in any of these businesses, frankly, Andrew, if I had a distribution advantage.
17:55Otherwise, I wouldn't get into any of these businesses.
17:58Okay. Fair. I think one of his distribution advantages is that he is so popular on Twitter also, and so that's why Peter Yang probably invited him on.
18:07That's why we're all following along. Um, he's also talked about his failed businesses like that Feather, which is the second from the top on the chart, is actually so low on the bottom of the chart that you can hardly see it. That was a tool for turning Notion pages into websites.
18:20He bought it, and essentially, the day after, Notion came out with the exact same features so you could build websites in Notion natively, and the thing just hanged. So but he talked about it publicly.
18:30He got attention. I like what you're saying there. Let's go on to the next one.
18:32I have I have a few more here. Alright. From starter story I built a mobile app that makes me $20,000
18:39per month. This is my app, the Cutcoach. It helps combat sport athletes cut weight for their competition or fight coming out.
18:47And the started building the Cutcoach app in June of last year using Cursor and ChatGPT. It took me around a month to build the MVP, and I gave it to my wrestling club to beta test after I finished building it.
19:01Alright. I wanted you to just see also how he's getting customers because Yeah. So for marketing, I would say to start off with organic posts, I got the ideas from social media pages that were in the same niche as mine, which I then adapted to my an example of the first video that I posted organically was a video where I showed what a UFC fighter looked like on a wake up and what they looked like after their wake up.
19:24I also added a call to action at the end of the video to give users an incentive to download my app. I noticed that the videos didn't get that many views, but I also noticed that the people that were viewing my posts were very high intent viewers. As I noticed around 10 to 15 downloads per day from posts that got anywhere from 200 to
19:45To 500 views. That's really killer. What do you think of this business?
19:49I think that it there's a probably unlimited amount of niche opportunities like this if you're willing to do organic social. And organic social meaning TikTok and Instagram.
20:01Mhmm. Um, if you're willing to do that and do what he did, you can find opportunities like this. Is this a business that gets to 2,000,000 a month?
20:09Probably not. Is it a business that can get to, like, 100 k, 50 k, all that? Probably.
20:14And I'm just sharing that criteria just so that if someone gets into these businesses, they understand this could do maybe 200 a month max.
20:23It probably can't do a lot more because there's just not that many people that need something like this or want something like this. At least, I'm imagining that. Right?
20:30Or it'd be really hard to get to them outside of these organic channels. So
20:36Give me a little bit about your background so people understand why you have such insights.
20:42For the last twenty plus years, I've just been doing product and marketing for online businesses, software and stuff like that, and built my own businesses.
20:52Funnily, you went quiet. You try to build stuff. You have no problem telling people that their that their ARR is not real.
20:58But when it comes time to telling me about, like, your background, suddenly, you went quiet. Alright. Let's go to the next one.
21:04Alright. This guy's got good money. Look.
21:06Later, we're at $100,000
21:08AR and more than 5,000 users. Meet Guy. It's a platform that takes that takes away the work of having to combine multiple tools in order to produce an AI video.
21:19So today, people use multiple tools from multiple platforms. They pay multiple subscriptions, and they need to have some technical knowledge to stitch everything together.
21:29Learner is aimed to
21:32AI explainer videos, next level. Here's the thing, though. Lou, what have you built it in?
21:37Then everything is managed very easily when you do white coding, especially with bass.
21:43So without having to give up on my studio completely, I was able to build something on the
21:50Let me pause. He's he's a developer. He's not developing.
21:53He's you you don't like the phrase vibe coding. I think he's vibe coding it.
21:57He's using base 44. Can we say anyone who's in base 44? That's, like, lovable.
22:01You're basically vibe coding.
22:03It's a better tool than the other ones quality wise from Really? Every yeah. We we've studied a bunch of these, and it actually produces better results than other ones right out of the box.
22:13Vibe coding or not, I think this you know, the other business was the AI video gen Mhmm. Business.
22:19This one is. I think one thing to keep in mind is that there's an arbitrage right now of people who know how to do these things and people who don't.
22:27Mhmm. And the arbitrage is basically people know who know how to do AI in this case, it's people who who know how to do AI generated video, which is not he's right.
22:36It's not easy. It's not easy to get it to be good. Or and great is a whole another problem.
22:42We actually completely built our type ahead video. It's on our YouTube. It fully AI.
22:50It's not on the homepage. It's not on the website, actually. It doesn't need to be.
22:53But it is on YouTube. You type in type ahead. You should find it.
22:55But that video is a minute long. It's fully AI. But it wasn't built in an hour.
23:00It took days and days because we wanted to use AI to to build it. Yeah, you should just type in type ahead, you should see it.
23:08And it's a minute long, it's a first Is this an idea? Yeah. And it's fully AI and Humans have been the exact same know the pain.
23:17Mhmm. I personally know the pain of trying to do what he's saying. And that pain is really hard.
23:23I mean, even in this case, we we didn't wanna pay for the stock audio for typing, so we actually made it with AI.
23:33Like, things like that, little things. But, like, this was, like, days and days of work. Like, this wasn't something we could do overnight, and you do one prompt and it'll come out like this or you give it a script.
23:43So I totally understand the business he's talking about and how hard it is to get that right. And if you get it right, then, of course, you might have that bright idea to go do this, you know, and build a product for it.
23:54But, again, I just wanna be specific here. There's a mental model where it's like things that took humans and cost hundreds to thousands of dollars are now able to be done for literally a fraction of the cost if you know how to use the tools.
24:11Okay. And so the interesting thing is, again, he's using Base 44 super easy. And look at how he's getting customers since you've said, look, the distribution is a big question.
24:22Is it gonna play? Did I tell you that I've ab coded this?
24:26I'll tell you. He's just using organic LinkedIn. He's not even buying any ads.
24:30Okay. I see. Alright.
24:32I like that too. Right? Yeah.
24:34No. It's great. I mean, the thing is he said studio, so I wonder if he had a video production or some kind of studio prior.
24:43And that's what gave him the idea and understanding of how to get these clients too.
24:48Okay. This is one of my favorite people. This is Greg Eisenberg doing an interview with someone who's doing $737
24:56a day. Look at this business. My Cloud Code directory, luxury restroom trailers, boring but awesome niche.
25:02People are spending a thousand dollars, $2,000 a day renting these out for their Get it. Weddings. You know, there's corporate events, film sets.
25:13Looks totally like the kind of thing that you can that, I'm gonna say, VibeCode, that you can use.
25:18Using Cloud Code and an open source GitHub repo called Crawler AI to handle any kind of data curation.
25:26So
25:28Okay. So fairly easy to build these things. These things, frankly, Heaton, I think have been easy to build even going back to the WordPress days.
25:34Right? Super simple, systematic. The the big question you're probably gonna ask is how is it getting These are top three ways to get distribution nowadays are ads, organic social,
25:43and SEO. Like, those are the big three.
25:46So choose your distribution. I made the mistake of not focusing on distribution first, and directories are a distribution first model. They kind of have an unfair advantage when it comes to SEO because if I create this directory or when I publish this directory, I kinda have this topical relevance where I have, like, a thousand pages all around luxury restroom trailers.
26:11Great idea. Right? You'd look for Austin Yeah.
26:13Luxury restroom.
26:15He could come up. Yeah. It's really smart.
26:17I I I think this is, again, a niche idea. He has an advantage of understanding it from some background or whatever it sounds like. This is good.
26:26This I would call vibe coded for sure because it's less software and more, like you said, things and directories and things we did. This is vibe codable. I think the hard part back in the day was getting the data.
26:37Now it's much easier to get the data. So now if I wanted to come up with another idea like this, the the luxury restroom trailers of something else, how do I find other topics, other things that I can set this up for?
26:50So how we found this back in the day with the directories was we did a bunch
26:55keyword research, right, on what people were typing into Google. And then we structured the directories around what people were typing in.
27:02Meaning, like, Google Trends type of thing? No. Like, using Ahrefs
27:07or or Semrush or one of those products or Ubersuggest and trying to figure out, like, what keywords are popular enough where building a directory around them makes sense.
27:19So for some reason, this luxury restroom trailer, it's really restroom trailer, think, or even like portable restrooms, things like that.
27:27I don't know the terms. Obviously, this person does. And so it's probably restroom trailer is my guess.
27:33And that build a directory entirely around that if it's like Austin restroom trailer or, you know, San Francisco restroom trailer or whatever.
27:43Although all of San Francisco is a restroom. Right? That's what they say.
27:47So yeah.
27:49These are big in Austin. Here we go. This final one, I don't think he's doing this,
27:57but his model is interesting, and I'd like to see what you think of this. Here's here's his model. Basically, what we're gonna do is we're gonna build out a web app tool, monetize it using ads.
28:06Now just to give you guys some examples of what I'm talking about, we're talking about things like mortgage calculators, typing tests, business name generators, other types of calculators. Those are super easy to create.
28:18Here's how much revenue he says you can Monthly traffic of, let's say, a million people. You're making between 3 to $6,000. It's
28:26Do you think, by the way, that that's possible having done this? Yeah. This is possible.
28:31This is totally possible. It just goes back to the research. Like, how are you gonna get the traffic for these things?
28:37If it's Google, then you gotta figure out how to get Google. And he's right. It's low balance and utility.
28:41Exactly what that summarizes exactly what it is, is very simple and well known, but most people don't understand this or do it, which is if someone comes from Google search, your job is to make sure they don't bounce. And that means you're providing utility for them so that the intent of the keyword they put in matches the landing page which then matches the next step that they take.
29:03Thus, they're not clicking back to Google. And if you have that, you can rank because your product's better than someone who should in Yeah.
29:11Correct.
29:13Here's how he's building it. I'd never Not need cleaning experience. All you basically do is use English to help guide the Horizon's AI builder.
29:21It's gonna understand the prompt that you submit. This is very similar to shipper where you just prompt it and Whatever you guys put. So as you can see, it understood what we are asking, and now it is starting to build out the foundation of our app.
29:34Over here on the right, if you wanna make any changes to what it's doing, uh, you can do all that right here just with written English. It might also ask you some questions or have you input certain information, and it's all done within here. Right?
29:45So you the thing about it, you're basically
29:47Okay. Basically, we we've got some Shannon with a how this business works, and he's he's vibe coding these here, there it is. The calculate your perfect mortgage payment.
29:57Good business model, easy to create, probably can spend three, five hours to make something like this, don't you think?
30:04Yeah. You can make it pretty fast. Yeah.
30:06I don't know about distribution. You still need to do things to get it promoted so Google will rank it and seize it. But Mhmm.
30:12Outside of that, yeah, the the hard part isn't the actual website.
30:17Of all these, which business do you think is the easiest one to create?
30:21The AI the AI the AI learning, the AI, you know, the one that's the AI, what is it, chief AI officer? I think that's actually the easiest.
30:30Oh, you're right. And he's the one who's, I think, making the most of all of these people. Think that's the easiest business right now.
30:36I think all the other ones are almost circling around the same thing. Not all of them, not the mortgage one, but anything that's AI is circling around this idea that there's AI arbitrage.
30:45AI arbitrage means you know how to do something with AI that other people don't. Whether it's the video gen or this stuff, it's all the same category in terms of framework of how to come up with ideas or what kind of ideas that you can come up with.
30:58Learn something that's valuable for companies, not individuals, but for companies around AI, and go try
31:05to go sell it to them. That's what he's doing. It's great.
31:08Why aren't you in this business? You've never been afraid of the consulting or or agency business. I I I
31:15think there's a most of my software will be in a business like this if it scales or as it scales because deploying AI is what companies are thinking through right now.
31:27I think the consulting and the training part, you you you still need distribution. I do free AI lessons every week. They're about thirty minutes of a lesson and then a bunch of q and a.
31:39This week's topic is about GitHub, and I'm doing a GitHub for non engineers sort of topic just to give people an idea of how you what you need to know about GitHub if you work at a company that uses GitHub even if you're not an engineer.
31:53Okay. And that's a way to understand AI better. So for me, it's more like trying to trying to trying to figure out what is worth helping people with that isn't gonna be this kinda rocky journey that doesn't need to be so rocky.
32:13And what I mean by that is new model comes out, you want your product or business to be better, not worse. Right? Things like that.
32:20So, yeah, I I think it's a good business. I I think more people should consider it.
32:25But again, to me, these agents or this . If you have skills on building a website and other people don't,
32:33you can essentially help them do it and get your get your money that way. Okay. So all those others make sense.
32:40This is the one. If I had to pick a business to start, this is the one. I agree with you.
32:44This does seem like the surest success,
32:47especially when you think about how he got his first customer to edit. So I just jumped on LinkedIn and started messaging some people, and eventually, you know, that that same day had a CEO, a local CEO that I didn't know really well, but we kind of had crossed paths a couple times. So it was I would call it a lukewarm lead.
33:03On LinkedIn.
33:04Alright, Ketan. This is exciting. Thanks for doing this with I'm looking forward to seeing more businesses.
33:09I wanna interview more entrepreneurs who build using these tools. Alright. Your point about this guy, John Cheney, being the person who has the easiest model is a good one.
33:17I interviewed multiple entrepreneurs who do it. In fact, on a video right here, you're gonna see me tell the stories of all those entrepreneurs, and you'll get to follow-up there. Go there next.
33:25See you there.
The Hook
The bait, then the rug-pull.
Eight founders. Eight different AI business models. One host who cheerleads for entrepreneurs and one who coaches them. By the end of 33 minutes, they've debated what's real, what's fake ARR, and which single business they'd both start from scratch today.
Named ideas worth stealing.
01:42concept
The Container Model for AI Ideation
Give AI a defined container — a specific problem and a target metric — then have it generate hundreds of ideas within that constraint. You pick the best one. This is different from asking AI to 'come up with business ideas' cold.
Steal forProduct discovery, feature brainstorming, any time you're optimizing within a known problem space
28:20concept
AI Arbitrage
People who know how to use AI to produce valuable outputs (code, video, automation, analysis) have a temporary but real knowledge gap over people who don't. That gap is a business. Either sell the output or sell the knowledge.
Steal forPositioning any AI-adjacent service or educational product
16:00model
Distribution-First Evaluation
Hiten's filter for every business: I'd only enter if I have a distribution advantage. The three current channels are ads, organic social, and SEO. Pick one you can win before building.
Steal forGo/no-go decisions on new business ideas
31:00concept
X Is the New Website
In the 2000s, knowing HTML when your neighbors didn't was a monetizable skill gap. Today, knowing how to prompt AI to build and produce outputs is the same gap — and you can sell into it on X/Twitter the way early webmasters sold into it.
Steal forPositioning for any AI-skills-based offer or content angle
12:40model
Forward-Deploy AI Officer Model
- Identify AI opportunities inside the client's business
- Implement the AI workflow or tool
- Charge recurring maintenance fee to keep it running
- Optionally: productize into private-instance software
The four-step model that turned a services engagement into a $4.5M claimed revenue business. Mirrors what Palantir and enterprise AI vendors do with forward-deployed engineers.
Steal forService-to-software transition planning, consulting offer design
How they asked for the click.
VERBAL ASK
33:10next-video
“In a video right here, you're gonna see me tell the stories of all those entrepreneurs, and you'll get to follow up there. Go there next.”
End-card with arrow pointing to a follow-up video. Andrew also mentions free weekly AI lessons as ongoing content hook.
MENTIONED ON CAMERA
Visual structure at a glance.
