Disclosure: SerpCtrl is an SEO agency. We build our own products (four currently in flight) and we operate a separate anchor-revenue agency line. We have no commercial relationship with any of the courses, builders, or tools named in this piece. We have used Cursor, Claude, and Supabase free tiers ourselves during builds; we are not an affiliate. The motivation for writing this is the same motivation that drives a senior SEO to publish a case study on a PBN: stop the people we care about from being separated from their money by a marketing surface that is engineered to hide its math.
TL;DR
The Instagram and TikTok claim, in its honest form: "With Cursor and Claude and a free Supabase tier, a non-engineer can ship a working web or mobile app for €200 to €1,500."
That claim is roughly true. It is the only part of the journey that costs €300.
Stacked on top of that true claim, the same accounts make a second claim: "Therefore you can build and earn six or seven figures of monthly revenue solo, in months not years, without needing distribution, support, compliance, or a failure portfolio of dead products behind you."
That claim is false at the base rate, and the people selling it know it, because the people selling it have no public Stripe-verified revenue from the apps they claim to have built, while the small number of indie hackers who do publish their numbers tell a completely different story: 5 to 11 years of work, 10 to 70 products shipped of which 80 to 95 percent failed, eventually one or two carrying the revenue load.
Here is the actual distribution, sourced and current to May 2026:
- 80.9% of Indie Hackers products earn under $500 MRR. Median MRR of revenue-reporting products: $303. (ScrapingFish analysis of Indie Hackers public data.)
- Only 4.8% of indie products earn over $10K MRR. $100K MRR puts you in the top 1%.
- 92% of micro-SaaS fail entirely. (RockingWeb 2025 study, n≈1,000.)
- 82% of indie SaaS never reach $1K MRR.
- 68% of mobile apps never reach 1,000 lifetime downloads.
- Only 17.2% of apps ever reach $1,000 in monthly revenue. (RevenueCat State of Subscription Apps 2024.) Median monthly app revenue at 12 months: under $50.
- Median time to $1K MRR for the indie SaaS founders who actually get there: 9 months. That is the survivors' time. The denominator includes the people who never get there.
This case study is in two parts. The first part honors the real builders by name with current revenue numbers, because they shipped, they published, they failed publicly, and they invite scrutiny. The second part maps the grift economy by name and by price, with the historical precedents that tell us where the wave goes from here.
The motivation, to repeat: this is not a case study against building. Building is the best thing a young person can do with their time in 2026. It is a case study against being scammed out of the money you would otherwise have spent on the actual work.
Part 1: The build cost claim, validated honestly
Before any debunking, let us validate the part of the claim that holds up. This is essential because every reader can confirm it with a credit card and a weekend.
The honest minimum stack to ship in 2026
- AI coding agent: Cursor Pro at $20/month, or Lovable at $25/month for 100 credits, or Bolt at $20/month for 10M tokens, or v0 at $30/month. Pick one. During an active 1-to-3-month build sprint, total spend is $20 to $90.
- LLM API for in-product use (if your product calls Claude or GPT in production): $20 to $100/month during heavy build, often $0 if you stay on a free tier during development.
- Database and backend: Supabase free tier handles 500MB database, 1GB storage, 50K monthly active users. Firebase free tier is comparable. $0 during build.
- Hosting: Vercel free tier is personal and non-commercial use only. The first hidden trap of the "$300 app" is that if you intend to charge money on day one, Vercel requires Pro at $20/month from the first transaction. Netlify and Cloudflare Pages have similar terms. The reel-bros do not mention this.
- Domain: Standard $10 to $15/year on Cloudflare, Namecheap, or Porkbun. Premium .com aftermarket: $200 to $5,000+.
- Stripe: $0 to start. Then 2.9% + €0.25 per EU transaction.
- App store certificates: Apple Developer Program $99/year. Google Play $25 one-off.
Honest total to ship a real working product in 2026: $200 to $500 for the weekend hero shipping web only. $1,000 to $1,500 if the build runs 2 to 3 months, includes a domain, paid AI subscriptions, an Apple cert, design assets, and a basic email service.
The Instagram claim survives this audit. Building has never been cheaper. A non-engineer with three months of focused work, a Cursor subscription, and a Supabase free tier can ship things that would have cost €50,000 to €200,000 of engineering time five years ago. This is a real and historically unprecedented democratization of building.
The claim collapses at the next step. Not at the build. At the operate. And then again at the earn.
Part 2: The actual revenue distribution
The reel-bros do not just claim that building is cheap. They claim that because building is cheap, the path from build to millions is short. The implicit promise is that 6-figure or 7-figure monthly revenue is reachable by a solo non-engineer within months.
This is where the math fails, and the math is publicly available to anyone who looks.
What 80.9% of indie products actually earn
ScrapingFish analyzed the public revenue listings on Indie Hackers, the largest indie SaaS community by public-revenue disclosure. The distribution:
- 80.9% of products earn under $500 MRR.
- Median MRR of revenue-reporting products: $303.
- Only 6.4% make over $5K MRR.
- Only 4.8% make over $10K MRR.
- Less than 1% make over $100K MRR.
These numbers are already optimistic, because they are conditioned on products that bother to report revenue publicly. The unreported median is closer to zero. Most products that get built do not get listed because they have nothing to list.
What the App Store and Play Store actually distribute
RevenueCat's State of Subscription Apps 2024 report:
- Only 17.2% of subscription apps ever reach $1,000 in monthly revenue.
- Median monthly revenue at 12 months after launch: under $50 USD.
Sensor Tower 2024 publisher distribution data:
- Small publishers earning under $1M/year account for 96.7% of App Store games publishers, and a tiny fraction of total App Store revenue.
- 68% of apps never cross 1,000 lifetime downloads.
What the failure rate looks like
- Median time to $1K MRR for indie SaaS founders who actually get there: 9 months. (Indie Hackers benchmarks, survivor sample.)
- 75th percentile: 12 months.
- 82% of micro-SaaS never reach $1K MRR. (Cross-sample average.)
- 92% of micro-SaaS fail entirely. (RockingWeb 2025, n≈1,000.)
For the people watching reels at home, the operative number is this: for every solo founder you see on Instagram claiming $50K MRR built in a weekend, the base rate says approximately 92 founders attempted the same thing in the same period and either failed entirely or earned under $1K MRR.
You are not seeing the 92. The reel format selects for the one survivor and shows that survivor's story as if it were the median. This is the textbook survivor-bias problem, identical to what happens when television shows you lottery winners but not the 99.9999% of ticket buyers who lost.
The point is not that the survivor is fake. The survivor is often real. The point is that the survivor's story is not the median story, and the reel framing is engineered to make you assume it is.
Part 3: The real builders, named and honored
Now the part that matters most. Because the answer to "the reel-bros are lying" is not "therefore indie SaaS is impossible." The answer is: the real builders exist, they shipped, they published their numbers, and what they tell you when you read their actual histories is the opposite of what the reel-bros tell you.
Every founder in this section publishes their MRR on a public dashboard, on Twitter/X, on a build-in-public blog, or via a Stripe-API-verified service like TrustMRR. Every one of them also publishes their failure portfolio. That dual disclosure (the wins and the losses) is the actual signal that distinguishes a real builder from a reel.
Pieter Levels (@levelsio)
- Current revenue: ~$3.1M ARR, ~$250K to $260K MRR as of November 2025. Public Stripe dashboard at levels.io.
- Products carrying the revenue: PhotoAI (~$132K/month), Remote OK (~$41K/month), Nomad List (membership), Interior AI, and a 2025 viral one-off FlyPieter that briefly hit $1M ARR in 17 days and is now archived.
- Years building: 11+ years. Started "12 startups in 12 months" in March 2014.
- Products shipped total: 70+ shipped. He publicly archives the failures on his site. Examples: Go Fucking Do It, Play My Inbox, YouTube Trends, dozens of others. He calls them out by name.
- Team: Solo. Zero employees. Publicly stated since 2014.
- Revenue split: 100% product. No course, no community, no upsell ladder. Famously hostile to the build-in-public-then-sell-a-course playbook.
The lesson from Levels is the cleanest in the dataset: 11 years, 70+ products, most failed, the survivors compound. A 22-year-old watching a reel and concluding "I can be Levels in a year" has misread Levels by an order of magnitude on both time and failure count.
Marc Lou (@marc_louvion)
- Current revenue: $1,032,000 earned across calendar 2025 (his own published year-in-review). January 2026 = $94,800. TrustMRR-verified MRR of $39,800 ($17K from TrustMRR product itself, $23K from DataFast, plus ShipFast and CodeFast on separate Stripe accounts).
- Years building: 6+ years. Started ~2019 while still in a corporate role.
- Products shipped total: 35 startups. 33 dead or near-zero by his own count. Two carry the revenue.
- Team: Solo founder. Hires contractors per project.
- Revenue split: Approximately 80% product (boilerplates, dev tools), 20% course (CodeFast at ~$23K MRR). Notably, the courses are about how to ship the products, not how to make money from courses. This is the legitimate end of the course spectrum.
Why he built TrustMRR: Because fake-Stripe-screenshot influencers had so polluted the "I make $XK MRR" claim that he built a Stripe-API-verified read-only dashboard tool specifically to make the verification trivial. TrustMRR exists because the grift exists. That is the most important fact in this entire case study.
Damon Chen (@damengchen)
- Current revenue: $109K MRR on Testimonial.to. Publicly posted on X in November 2025. Also runs PDF.ai.
- Years building: 5+ years on Testimonial.to (founded December 2020). Multiple failed products before that.
- Products shipped total: A handful before Testimonial.to, all dead. Quit his job in March 2021 when Testimonial.to was at $1K MRR.
- Team: Solo founder for the first few years; small team now.
- Revenue split: 100% product (B2B SaaS).
The Damon Chen story is the closest fit to the "reasonable hope" version of indie SaaS: a real software engineer, multiple shots, finally hit a product-market-fit win at age 30+, took 5 years from that win to $109K MRR. Not a reel. A career.
Tony Dinh (@tdinh_me)
- Current revenue: TypingMind $130-160K/month, DevUtils $5K/month (peaked $20K), various smaller products. Crossed $1M cumulative in 2025, currently working ~20 hours/week.
- Years building: 6+ years building publicly.
- Products shipped total: 10+ shipped including BlackMagic (peaked $14K/month, now lower), Xnapper, plus, in his own words, "a ton of failed products."
- Team: Solo.
- Revenue split: 100% product. Sells a $40 book "My Indie Book" but that generates approximately $518/month, marginal.
Daniel Vassallo (@dvassallo)
- 2025 exit: Sold Small Bets community to Gumroad in April 2025 for $3.6M (50% cash, 50% Gumroad stock).
- Pre-sale revenue (Nov 2021 to Oct 2023): Small Bets generated $824K revenue, $600K profit.
- Years building: 5 years since leaving Amazon in 2020.
- Notable acknowledged failures: Userbase SaaS ($10K/year, he calls it his failed bet), Profit and Loss community ($5/month, died), Small Bets Deals ($31K invested, failed experiment).
- Why he matters here: He publishes the failures with the wins. He is the rare community-product operator who refuses to pretend that the playbook he sold was a sure thing for the buyer. The community itself is named "Small Bets" because the operating thesis is that most bets lose.
Justin Welsh (@thejustinwelsh)
- Cumulative revenue: $10M+ by 2025. 2024 = $4.15M. 2023 = $2M+.
- Revenue sources: The LinkedIn Operating System course ($150) + the Content OS, the Creator MBA, newsletter sponsorships ($5K/issue × ~50 issues = $250K).
- Years building: 6 years as solopreneur (started 2019).
- Why he is in this list: Because he is a course-seller who is the legitimate end of the course-seller spectrum, and the contrast matters. He sells digital products, he discloses his numbers, his courses have indexed pricing pages, he does not pretend to be primarily a SaaS builder, and his outputs (the LinkedIn Operating System) are practical, testable, and reviewed positively by buyers who include him in their case histories without being affiliates.
The Welsh case is the necessary calibration. The argument is not "all courses are grift." Welsh sells courses and is not a grift. The argument is that there is a recognizable difference between courses with indexed pricing, indexed reviews, and indexed delivery, versus reel-only DM-to-buy cohorts at $5K to $30K with unverifiable revenue claims on the seller side. Welsh is the former.
Jon Yongfook (Bannerbear)
- Current revenue: ~$50K MRR. Public open-startup dashboard.
- Years building: 8+ years solo on Bannerbear.
- Product: Visual content API. B2B SaaS, paid plans only.
- Revenue split: 100% product.
The pattern across every named winner
Look at the list. Then look at the years column. Then look at the products-shipped column.
1. Builder: Pieter Levels
Years: 11+
Products shipped: 70+
Most failed? - Yes (most)
Public Stripe? - Yes, live dashboard
2. Builder: Marc Lou
Years: 6+
Products shipped: 35
Most failed? - Yes (33 of 35)
Public Stripe? - Yes (TrustMRR)
3. Builder: Damon Chen
Years: 5+
Products shipped: Many
Most failed? - Yes (all before Testimonial.to)
Public Stripe? - Yes (public posts)
4. Builder: Tony Dinh
Years: 6+
Products shipped: 10+
Most failed? - Yes ("a ton")
Public Stripe? - Yes (public posts)
5. Builder: Daniel Vassallo
Years: 5+
Products shipped: Multiple
Most failed? - Yes (acknowledged)
Public Stripe? - Yes (public sale price)
6. Builder: Justin Welsh
Years: 6+
Products shipped: 4-5
Most failed? - Some
Public Stripe? - Yes (annual disclosure)
7. Builder: Jon Yongfook
Years: 8+
Products shipped: Several
Most failed? - Yes (Bannerbear is the win)
Public Stripe? - Yes (open dashboard)
The shortest time-to-success in the entire named-winner list is 5 years. The shortest product-count-to-success is a handful, with most dead. Every single named winner publishes their numbers. Every single one talks publicly about the failures.
Now compare that pattern to what the reel-bro says: "I built this app last month. I'm doing $50K MRR. Here is my course on how."
You do not need to be a senior SEO or a forensic auditor to spot the inversion. You only need to read the column headers.
Part 4: The CAC math that kills the dream before it starts
Let us assume, for the sake of argument, that you have built a real product. The build cost is real. The product works. The pricing page is live. Stripe is connected. You posted the launch on Twitter.
What happens next is the part the reel format leaves on the cutting-room floor.
B2B SaaS CAC, 2025 benchmarks
- Median B2B SaaS CAC: $1,200. (Maxio 2025 SaaS Benchmarks Report.)
- CAC rose 14% year-over-year.
- Healthy LTV:CAC ratio: 3:1. Sub-$2M ARR companies can tolerate 2:1 or 2.5:1 short-term.
- Industry rule of thumb: "$2.00 of sales and marketing to acquire $1.00 of new ARR." (Benchmarkit.)
If you build a B2B SaaS at $50/month with a 24-month average customer lifetime, your LTV is $1,200. Your CAC ceiling at 1:1 is $1,200, at 3:1 is $400. Welcome to the reason most B2B indie SaaS founders go bankrupt or burn out in year two: the CAC has to come from somewhere, and "somewhere" is either paid ads (impossible to make work at $50/month price points) or 18 months of organic content (which is unpaid founder labor).
B2C SaaS CAC reality
At a $7/month consumer subscription with 10% monthly churn (typical), the LTV is approximately $28 ($7 / 0.10 for the churn-adjusted simple LTV after collection fees). Your maximum sustainable CAC at 3:1 is $2.80.
Two dollars and eighty cents to acquire a customer.
The current Meta CPM (cost per thousand impressions) is $14.19. The average B2C SaaS conversion rate from click to paid is around 2 to 3%. Meta SaaS CAC averages $55.21.
If your CAC ceiling is $2.80 and the cheapest paid acquisition channel delivers customers at $55, you cannot use paid acquisition. You have to acquire customers organically. Which means SEO content, social, viral, community, or referrals. Which means 18 to 36 months of compounding unpaid labor.
This is why the median time to $1K MRR for indie SaaS founders who get there is 9 months, and why the people who do not get there in 12 months mostly stop trying.
The "free" organic-marketing labor cost
If you do go the organic route, the cost is not zero. It is your time. The Indie Hackers benchmarks survey shows that the 9-month median to $1K MRR is roughly 20 to 40 hours/week of founder effort during that period, mostly on marketing, support, and iteration rather than coding.
At any realistic opportunity cost ($50 to $150/hour for a skilled builder):
- 20 hours/week × 36 weeks × $50/hour = $36,000 of unpaid labor.
- 40 hours/week × 36 weeks × $150/hour = $216,000 of unpaid labor.
That labor is real. It is the part of the iceberg under the waterline that the $300 claim hides.
The reel-bro never says "$300 to build, then $36,000 to $216,000 of unpaid distribution work." They say "$300 to build, $50K MRR in three months, buy my course."
The compounding math problem
To make $1M in revenue at a $100 LTV (a healthy SaaS LTV for a small product), you need 10,000 paying customers. If you can acquire them at $50 CAC (better than the Meta SaaS average), that costs $500,000 in CAC alone. At $300 CAC (more realistic for B2B), it costs $3,000,000.
The reel claim is that you can do this with $300 and no team. The math is that you would need to find 10,000 customers willing to pay you, organically, with zero paid acquisition, which has been done (Levels has done it, Marc Lou has done it, Chen has done it), and which takes 5 to 11 years when it works.
Part 5: The real operating cost to $10K MRR
Beyond CAC, there is the silent budget eater: the cost of actually running the thing once you have customers.
EU VAT and US sales tax
If you sell to EU consumers from anywhere in the world, the first sale to an EU customer triggers VAT registration. There is no threshold for non-EU sellers. (The threshold exists only for intra-EU sellers, and only on B2B.)
Solo founders almost universally outsource this to a Merchant of Record (MoR) like Paddle, Lemon Squeezy, or Polar, which handles VAT registration and remittance across 27 EU countries in exchange for approximately 5% + €0.50 per transaction. Real-world MoR stack costs, with currency conversion and payout fees, reach 7.9% to 18% of revenue. (Lemon Squeezy users widely report 10 to 14% effective.)
For US: Stripe Tax at 0.5% of taxable transactions, capped at $0.50/txn. Manageable but not free.
Hosting at scale
- Supabase Pro at $25/month kicks in around 1,000 active users.
- Vercel Pro at $20/month is required from day one of any commercial app (the free-tier-trap mentioned earlier).
- Bandwidth at $0.09 per GB is the silent overrun. 100,000 monthly active users on a media-heavy product can land you a $200/month bandwidth bill from Cloudflare or Vercel.
- Typical indie SaaS at $1K to $5K MRR pays $55 to $95/month in baseline hosting. At $10K MRR, $150 to $400/month.
Email infrastructure
- Resend free up to 3,000 emails/month, then $20/month for 50,000.
- Postmark $15/month entry tier.
- Loops ~$49/month at 5,000 contacts.
- Mailchimp $20/month Standard at 500 contacts. $135/month at 10,000. Premium tier balloons further.
- Realistic indie stack: $40 to $100/month at the $1K to $10K MRR range.
Analytics and customer support
- PostHog free tier is generous (1M events, 5K replays). Paid jumps to $450/month at 50M events. For an indie SaaS this is usually the only viable analytics tool.
- Mixpanel free 1M events. Amplitude paid starts $1,000/month.
- Support time: rule of thumb is approximately 1 hour per 100 customers per week for self-serve SaaS, 2 to 3x that for B2B with onboarding. That is 40 to 200 hours/month at $1K to $10K MRR.
Refund and chargeback drag
- Apple App Store / Google Play refund rates: blended 6 to 7%. Education 4.86%, health & fitness 4.71%. (Business of Apps 2026 data.) Apple takes 15% (Small Business Program under $1M) to 30%.
- Stripe dispute thresholds: Visa is lowering from 2.2% to ~0.9% in April 2026. Stripe internally treats above 0.75% as high-risk. Hit 1%+ disputes and Stripe holds reserves, hikes fees, or terminates.
- Effective economics on a €15/month B2C app via Apple's App Store: €15 charged, minus 15% Apple, minus 5% refund risk, minus churn = ~€10 to €11 retained per customer per month at year one.
The total real cash outlay to $10K MRR
1. Item: Build cost (one-off)
Range: $200 - $1,500
2. Item: Domain (annual)
Range: $10 - $200
3. Item: Hosting + email + analytics (12 months)
Range: $1,500 - $4,500
4. Item: Legal templates, accountant, basic compliance
Range: $500 - $3,000
5. Item: MoR / Stripe / Apple haircut on revenue
Range: 8% - 25% of revenue
6. Item: Optional paid acquisition to accelerate
Range: $0 - $30,000
7. Item: Refunds and chargebacks
Range: 3% - 7% of revenue
Realistic cash outlay to reach $10K MRR: $5,000 to $50,000.
Realistic unpaid founder labor to reach $10K MRR: $30,000 to $200,000 at opportunity-cost wages.
The $300 claim describes one item on the first line of that table. The reel-bros price the entire journey at that line item.
Part 6: The grift economy, mapped by name and by price
This is the part where it matters to be specific, because vague critique is useless. Naming is how the diagnostic gets learned.
The upsell ladder (industry-standard)
Every course-grift in this category runs the same four-rung architecture:
- Free content: reels, TikToks, YouTube, podcast episodes. Lead generation.
- Low-ticket entry: $27/month community, $97 to $497 introductory course. The first credit-card swipe.
- Mid-ticket cohort: $2,000 to $5,000 for a "live cohort" or "intensive."
- High-ticket mastermind: $10,000 to $30,000 for "1:1 access," "Discord inner circle," or "AI agency accelerator."
Each rung exists to upsell the buyer from the previous rung. The reels at rung 1 are not the product. The product is the ladder.
Named accounts and current pricing
These are publicly indexed prices as of May 2026. The prices are not the criticism. The criticism is in the structure and in what is not publicly indexed.
- Riley Brown (@rileybrown_ai) - The #1 "vibe coding" YouTube account. Sells a free "Vibecoding 101" introductory course on Maven, with a paid flagship course pre-launched for January 2026. Public price for the flagship course: not indexed. Waitlist-gated. The absence of indexable pricing is the diagnostic signal: indexed pricing invites comparison; DM-to-buy or waitlist-only invites information asymmetry.
- Greg Isenberg (Late Checkout) - Community Empire membership at $99/month, plus a Maven cohort "Community College." Has a real prior track record (sold Islands, advised on multiple companies). The current revenue mix appears to be primarily community membership + cohort + advisory rather than products. Textbook four-rung ladder: podcast → newsletter → $99/month → cohort → 1:1 paid call. Legitimate in the sense that the products exist and are indexed; the buyer should know they are buying community/teaching, not a SaaS playbook with guaranteed outcomes.
- Dan Koe - Modern Mastery community at $27/month; "How To Productize Yourself" at $159; a "3-Day Intensive" periodically sold without public pricing. Creator-economy operator, not specifically an AI app grifter, but the canonical four-rung ladder.
- Carson Reed - AI Agency Mastermind - $10,000. Flagged by multiple independent reviewers (Ippei, others) as scam-patterned: "earnings proof" comes from course sales presented as agency revenue, sock-puppet positive reviews, a 6-month access window too short to actually build the claimed business. Canonical $10K-tier grift.
- Andrew Tate - The Real World (formerly Hustlers University) - $99.99/month, doubled from the original $49.99. Affiliate program was shut down by Tate in August after Meta banned the campaigns and Stripe pulled processing for HU's predecessor. Rebranded continues. Not specifically about app building, but the same architecture and the same playbook that the AI-app grifters borrowed.
- Marc Lou - ShipFast at $199 launch (now ~$54K MRR per public reporting), . Indexed pricing, real product, public verification. , because the fake-screenshot economy had polluted his category. . He is also not above legitimate criticism - SaaS founders have publicly disputed his "$100K in 4 months" framing as marketing-led rather than literally accurate, and ShipFast had a paywall-bypass security incident he publicly acknowledged. The point is: the disputes are themselves indexed and public. That is the difference.
The fake Stripe-screenshot economy
The grift is dense enough that an industry exists for the fake screenshots themselves:
- fakemrr.app and fake-mrr.com are operational tools that generate convincing Stripe dashboard screenshots with any MRR number you input.
- Kyle Gawley (founder of Userloop) tweeted about the existence of a "fake Stripe MRR generator" he built as satire in 2024; the tool is currently the #3 indexed page for that search term.
- TrustMRR (Marc Lou) and FakeMRR.com (pierbapt) are the community countermeasures: read-only Stripe API verification for the former, and a public database of unverified-claim startups for the latter.
The existence of fake-MRR tools as a productized category, with countermeasures built and indexed, is the strongest single signal that the screenshot-claim economy is hollow at scale. If verification were cheap and the claims were real, neither tool would have a market.
The affiliate-disclosure problem
Multiple "vibe coding" influencers earn from affiliate referrals to the tools they push in tutorials. Public referral programs:
- Lovable: 20% recurring commission for 12 months, 60-day cookie.
- Anthropic Referral Partner Program: launched March 26, 2025, case-by-case negotiated rates for enterprise-tier influencers.
- Cursor: no public referral program per forum threads, but partnership arrangements at the enterprise tier exist quietly.
- v0/Vercel, Bolt: similar undisclosed partner programs.
When an influencer says "I built this with Cursor and it changed my life," and Cursor (or its peer) is paying that influencer a recurring commission on every signup, that fact should be disclosed under FTC Endorsement Guides and EU advertising-transparency directives. It almost never is. The disclosure absence is the second indexable diagnostic signal.
Karpathy's own concession
Andrej Karpathy coined the term "vibe coding" in a viral February 2025 tweet describing the experience of coding by intuition with AI. In a follow-up thread in March 2025, Karpathy himself acknowledged that the term "kickstarted a new wave of grifting". He engaged with the meme thread documenting this approvingly.
When the originator of the term concedes the category has become a grift wave within 30 days of coining the term, that is dispositive evidence about where the category went.
Part 7: The historical wave precedents
This is the part that frames the prediction. The pattern is genuinely cyclic. Every 3 to 5 years, a new technological wave creates a real democratization of building or earning, the real opportunity attracts the first wave of legitimate operators, and within 18 to 36 months a parasitic course economy attaches to the wave and runs the same playbook until the wave collapses or the payment processors pull out.
1. Wave: Forex retail trading
Real opportunity (years): 2005 - 2012
Parasitic course peak: 2008 - 2012
End event: MyFXBook regulation, EU CFD restrictions 2018
2. Wave: Crypto / ICO
Real opportunity (years): 2013 - 2017
Parasitic course peak: 2017 - 2018
End event: 2018 ICO collapse, SEC enforcement
3. Wave: Dropshipping / Shopify
Real opportunity (years): 2015 - 2022
Parasitic course peak: 2018 - 2022
End event: iOS 14.5 ATT (Apr 2021) destroyed Meta ad ROAS for ecom
4. Wave: NFTs
Real opportunity (years): 2020 - 2021
Parasitic course peak: 2021 - 2022
End event: 2022 floor collapse
5. Wave: OnlyFans agencies
Real opportunity (years): 2020 - 2024
Parasitic course peak: 2022 - 2024
End event: Vice/Prism Reports exposure 2024, platform crackdowns
6. Wave: Andrew Tate Hustlers University
Real opportunity (years): 2021 - 2023
Parasitic course peak: 2022 - 2023
End event: Meta ban + Stripe pull, August 2023
7. Wave: AI / vibe coding
Real opportunity (years): 2023 - active
Parasitic course peak: 2025 - active
End event: Predicted: Aug 2026 - Feb 2028 based on 18-36 month wave median
The AI/vibe-coding wave started in earnest with Karpathy's February 2025 tweet and has been at peak hype since. The 18-month minimum precedent puts the first major collapse window at August 2026, plausibly triggered by either a payment processor pulling a major mastermind, a high-profile cohort buyer suing for misrepresentation, or a regulator (ASCI in India has already warned about fake AI training claims) issuing a category-level advisory.
The 36-month maximum puts the outer bound at February 2028.
This is not a prediction of doom. The underlying technology (LLM-augmented coding) is real and permanent. The wave that will collapse is the parasitic course economy attached to it, not the technology. Levels and Marc Lou and Damon Chen will still be shipping in 2028. The $10K mastermind operators will not.
Part 8: The 15-minute diagnostic for spotting reel-bro vs real builder
Apply to any "I built this app and now earn $X MRR" claim you encounter. If three or more of the green flags are absent and any of the red flags are present, the claim should be treated as marketing not as evidence.
Green flags
- Stripe-API-verified revenue (live dashboard, TrustMRR, or similar), not screenshots.
- Public failure portfolio. The builder names their dead products with the same specificity as their winners.
- Multi-year timeline. The builder shows the work going back 3+ years, with the slow ramp visible.
- Indexed pricing pages for any courses or products. No DM-to-buy gates. No waitlist-only mystery pricing.
- Indexed third-party reviews on platforms with anti-gaming controls (Indie Hackers, Hacker News, Product Hunt, G2 if applicable).
- Affiliate disclosure when recommending tools.
- Engagement with critics. Real builders engage with critique. Grift accounts block.
- Numbers that survive simple verification. If they claim 10,000 users, the product's public usage signals (subreddit size, Twitter mentions, App Store reviews, web traffic estimates from Similarweb) should be consistent with 10,000 users.
Red flags
- Course revenue exceeds product revenue. If they make more from teaching how to do the thing than from doing the thing, the course is the product and the "thing" is the marketing.
- DM-to-buy or waitlist-only pricing on any cohort/mastermind.
- Stripe screenshots without verification, especially zoomed-in single-month snapshots.
- Time-collapsed claims. "I built this last month and now make $50K." Median time to $1K MRR for survivors is 9 months. Time-collapsed claims of $10K+ MRR in under 3 months are nearly always selection bias from a one-hit viral moment that is being represented as a repeatable system.
- "I built 10 apps in 10 weeks" portfolios where most apps are empty Stripe landing pages with no users.
- The brother-of-the-CEO testimonial pattern (covered in our BabyLoveGrowth case study). Marquee testimonials from related entities, not arm's-length customers.
- Aggressive affiliate program with no qualification gate. When the recurring commission is high and anyone can sign up, the review economy gets gamed.
- Cohort/mastermind pricing above $5K with no public testimonials from indexable buyers. Real $10K cohorts get reviewed, criticised, and discussed in public.
This 15-minute diagnostic catches the overwhelming majority of grifts in this category. Apply it before you swipe the card.
Part 9: What is actually realistic, and worth doing anyway
This case study would be worse than the grift it critiques if it left the reader with the conclusion "do not bother building." That is not the lesson. The lesson is the opposite. The lesson is: build with accurate expectations, and you will keep building long enough for one of your bets to land.
Realistic expectations, calibrated against the named-winner cohort:
- 6 to 18 months to your first $1K MRR if you are a skilled builder with an existing audience or working in a B2B niche you understand. Longer if you are still learning.
- 2 to 4 years to $10K MRR, with at least 3 to 10 dead products behind you.
- $100K+ MRR requires either luck (a viral one-off, like FlyPieter's $1M in 17 days), exceptional skill compounded over a decade (Levels, Marc Lou, Damon Chen), or a team. It is not a solo three-month achievement, ever, for anyone. The reel-bros' $100K MRR claims are almost always either fabricated, course revenue mislabelled as product revenue, or one-month spikes presented as steady-state.
The right framing is what Daniel Vassallo named his community: Small Bets. Most bets lose. You are not trying to hit one big win; you are trying to survive long enough to take enough small bets that one of them eventually catches.
The right cost to bring to this game is small enough that 35 failures still leaves you solvent. That is Marc Lou's actual playbook, not the marketing version of it. He shipped 35. Two paid off. The 33 cost him a few hundred dollars each in domain, hosting, and AI subscriptions. He is in his sixth year, and now he runs at $1M+ per year. The math works because each individual bet was cheap, not because any individual bet was big.
Part 10: The motivation, restated
The reel-bro economy works because aspiration is the most valuable raw material in the creator economy and there is no shortage of it. The people who buy the $497 course or the $5,000 cohort are not stupid. They are skilled, motivated, and underserved. They are willing to spend, because spending is doing-something, and doing-something feels better than the harder thing, which is building for years with a high failure rate while reading public dashboards from people who have already done the time.
The $497 course transfers their money to someone whose primary product is the course. The $5,000 cohort transfers it to someone whose primary product is the cohort. The $30,000 mastermind transfers it to someone whose primary product is the mastermind. None of these transfers move the buyer measurably closer to their stated goal.
The motivation for writing this case study is not to be cynical about building, or to discourage anyone from trying. It is the opposite. The $30,000 a young builder spends on three cohorts and a mastermind in their first year is enough to fund 60 to 300 shipped products. That is roughly Marc Lou's first three years of work measured in raw build cost. The money would have been better spent building, failing, building again, and reading Levels' public history.
If you are reading this and you have already bought one of those courses, you are not stupid. The market is engineered to extract from you. Stop buying. Start shipping. Use the public dashboards of Levels, Marc Lou, Damon Chen, Tony Dinh, Daniel Vassallo, Jon Yongfook, and Justin Welsh as the calibration of what success in this category actually looks like and how long it actually takes.
The $300 build claim is the only honest number in the reel. Take that number, multiply it by 30 shipped attempts, give it 5 to 10 years, and you have a path that has worked for the people who actually got there. There is no faster version. Anyone selling you a faster version is selling the course.
Methodology and disclosure
This case study uses publicly indexable revenue data, Indie Hackers leaderboard statistics, RevenueCat State of Subscription Apps 2024 report, Sensor Tower 2024 publisher distribution, Maxio 2025 SaaS Benchmarks, Benchmarkit 2025 SaaS Performance Metrics, Stripe's published dispute and fee policies, and the named founders' own public revenue disclosures via their personal sites, X/Twitter posts, TrustMRR verification, and Stripe dashboards where public. No private data was used.
Course prices are taken from the named operators' public pricing pages, Maven course pages, or community membership signup pages as visible in May 2026. Where pricing is gated (Riley Brown's flagship course, certain mastermind tiers), the absence of public pricing is reported as the relevant fact.
Affiliate programs are taken from the tool vendors' public affiliate or partner program pages or from indexed forum discussions where vendors do not publish program terms (Cursor).
For any factual error: admin@serpctrl.lv. Corrections will be made publicly.
SerpCtrl operates under SIA Cyber Unicorn (Latvian registration 40203002129). We provide SEO audit, monitoring, and managed services. We are also currently building four products of our own (Stoa, Piespēle, ConferencesCost, QuestBuddy) and we have shipped many things that did not work over the past ten years. We have no commercial relationship with any of the named courses, products, builders, or tools in this piece, except that we use Cursor and Claude in our own building work and have used Supabase and Vercel free tiers, without affiliate compensation.
For SEO and product strategy services that help founders separate signal from noise on the way to actually shipping something: serpctrl.lv