AI Side Hustles Misleading — the headline number is real but unrepresentativ

By the Income Reality Check Editorial Team · Updated 2026-07-10 · 7 min read

RizzGPT to Cal AI: the $1M-a-month app story the video leaves half-told

Verdict: Misleading — the headline number is real but unrepresentative. The apps genuinely earn seven figures; the “any non-coder can do this with ChatGPT” packaging quietly drops the co-founder, the engineer, and the paid-distribution engine.

Vaibhav Sisinty’s video opens with a founder pulling in “almost 9 crore rupees a month” — roughly $1 million — after 18 months, having supposedly never written a line of code. The founder is Blake Anderson, and the apps are real: RizzGPT, Umax, and Cal AI. The money is real too. What’s misleading is the shape of the story, which files the sharp edges off a business that took partners, capital, and a distribution playbook most viewers won’t have.

What the video actually claims

The pitch is a three-move “playbook” you can supposedly run on your own idea this weekend. Move one: find a cultural obsession people already care about (dating, looks, calories). Move two: build fast by treating an AI model like a junior engineer and “shipping the ugly version.” Move three: treat distribution as the product — pay micro-creators $50 a post to seed the app on TikTok.

As proof, the video cites specific revenue at each step. RizzGPT hit “around $80,000 a month” after crossing 200,000 downloads in a week, and “still generates close to $200,000 every month.” Umax earned “$100,000 in month one,” then $200,000, then “half a million dollars regularly,” past 10 million Play Store downloads. Cal AI is the headline act — the app behind the million-a-month figure.

The framing throughout is that Blake “just figured out how to sit in front of ChatGPT and ask it the right things.” He’s “not even an engineer.” That single sentence is doing a lot of work.

What the method actually requires

Start with the co-founder the video mostly erases. CNBC and TechCrunch both report Cal AI was built by a team, not a lone prompt-whisperer. Zach Yadegari — who began coding in middle school, built a gaming site called Totally Science, and sold it for $100,000 at 16 — is a co-founder, alongside Henry Langmack and a COO, Jake Castillo, who runs influencer marketing (CNBC; TechCrunch). The video itself admits, in passing, that “his brother helped by doing a little engineering back end.” A little engineering is still engineering. The person who can direct an AI model to fix an exposed API key is not the same person who doesn’t know one exists.

Then there’s the money going out. The $50-a-creator anecdote sounds almost free, but the actual model the video describes is industrial: DM about 100 creators to get 10 replies and maybe 3 conversions, then reinvest every time 1,000 views earns more than it costs to buy. That’s paid acquisition with a full-time operator (Castillo) behind it. It works only if your unit economics already work — and finding out whether they do costs real money up front.

The platforms take their cut before you see a rupee or a dollar. Apple’s standard commission is 30% on subscriptions in year one, dropping to 15% under the Small Business Program for developers under $1 million in annual proceeds (Apple Developer). Google Play charges a similar 15% on the first $1 million (Google Play Help). Cross the million-a-year line and Apple’s cut on much of your revenue jumps back to 30%. CNBC’s reporting reflects this: Cal AI’s ~$1.4 million monthly is described as gross profit after the store cuts — meaning top-line revenue is higher, and so are the costs behind it.

Here’s a rough monthly picture for a subscription app at scale, to show where the headline number goes:

Every image Cal AI analyzes runs through a paid vision model. At millions of downloads and repeat daily use, those API calls are a real recurring bill, not a rounding error — which is exactly why “ship the ugly version” only survives if something is monetizing fast enough to cover the meter.

Is the million-a-month figure even representative?

No. It’s a top-of-the-market outcome quoted as if it were a repeatable recipe.

The mobile app economy is brutally top-heavy. Independent industry data consistently finds that the vast majority of apps earn almost nothing: the median subscription app makes under $50 a month after a year, only around 17% ever reach $1,000 in monthly revenue, and roughly 3.5% clear $10,000 — the rough floor for going full-time. The top 1% of apps capture the overwhelming majority of all store revenue. Blake Anderson is in that 1%. Building your plan around his number is like planning your NBA career around LeBron’s contract.

The video also compresses timing and luck into “a playbook.” Umax rode GPT-4 Vision’s launch window — a brief moment when face-rating was suddenly cheap to build and nobody else had shipped it. That window is a real, repeatable idea (watch for a new capability, build the obvious app first). It is not a repeatable result, because by definition only the early movers get the empty category.

Who actually wins this game

Early movers with a technical partner and a distribution operator. That’s the honest profile. Blake spotted waves early — but he also had a brother doing backend work, a co-founder who’d already sold a company, and a teammate whose whole job was creator marketing. The video’s own lesson, repeated twice, is “he won on distribution, not technology.” Correct. And distribution at this scale is a paid, staffed, tested function — not a $100 TikTok hack you run once.

The people this doesn’t work for are the ones the “no code needed” line is aimed at: a solo builder with no engineering fallback, no budget to lose testing creator hooks, and no way to spot a fresh API capability before the crowd does. AI removes the coding bottleneck. It does not remove the distribution bottleneck, and it does not hand you a co-founder.

What you’d realistically earn

If you’re a beginner following this exact method with modest capital, plan for $0 for the first several months while you build, ship, and test distribution — with a real chance of never clearing your API and ad-testing costs. The apps that break out usually do so because their creators kept firing until one hit, absorbing the losses on the misses. A more grounded target for a competent solo builder who lands a genuinely useful app and works distribution hard: maybe a few hundred to a few thousand dollars a month after a year, minus platform cuts and API bills. Seven figures a month is not a target. It’s a lottery outcome with a good story attached.

One more thing worth knowing if you’re in the U.S. The FTC has proposed rules specifically to curb deceptive “money-making opportunity” earnings claims, requiring sellers to have real, empirical substantiation — not anecdotes — for income figures (FTC). The video isn’t selling you a course outright, but it does funnel viewers toward a paid “Claude code masterclass” and a WhatsApp community using one man’s outlier earnings as the hook. Treat headline income numbers as marketing until someone shows you the typical result.

Who this is (and isn’t) for

This approach fits someone who can already build or has a partner who can, has $500–$2,000 they can afford to burn testing creator distribution, and can commit real weekly hours to shipping and iterating for months with no guaranteed payout. It also rewards people genuinely plugged into an online subculture, because that’s where the unmet-need signal lives. It’s a poor fit for anyone who needs income soon, can’t absorb losses, or is taking “no skills needed” literally — that phrase is the part of this pitch that’s least true. (If you want to test the water-line, see how you’d start a one-person business with Claude AI in 30 days and five realistic ways to make money in 2026 with AI.)

What to remember

Blake Anderson’s numbers are real, and his core insight — that distribution beats polish — is genuinely good. The misleading part is the packaging: a staffed, funded, early-mover team’s outlier result sold as a solo, no-code, weekend playbook. Copy the thinking. Just don’t budget around the headline.