AI Side Hustles Half-true — works only if you do the unspoken work

By the Income Reality Check Editorial Team · Updated 2026-06-30 · 7 min read

Claude AI side hustles that beat a 9-to-5? The catch is the audience

Verdict: Half-true — works only if you do the unspoken work. The Claude tooling is real, but every success story in the video is powered by an audience the pitch treats as the easy part.

Shane Hummus opens “5 Claude AI Side Hustles That Pay More Than a 9–5” by promising you’ll know how to start each one this week, and he name-drops eye-watering numbers along the way: a realtor whose one video beat his entire channel history, a coach who made $28,000 in a month, a KDP publisher who made over $500,000 in a single month. The five hustles are real categories of business. So is Claude’s usefulness inside them. But watch closely and all five quietly collapse into one requirement the video rushes past — building an audience that knows and trusts you — and that’s the part Claude can’t do for you.

What the video actually claims

Hummus lays out five plays. The “boring gold mine” is running a normal local business (real estate, plumbing, cleaning) and letting Claude handle the marketing, emails and follow-ups. The “silent bookstore” is publishing on Amazon KDP, with Claude drafting whole books. The “hidden expert” is coaching or consulting on a skill you already have. The “invisible workforce” is a done-for-you service agency where Claude does the production. And the “audience machine” is a niche YouTube channel — which, by his own admission near the end, is “the secret hiding under all five.”

To his credit, he says it plainly: “Every single person that I just showed you… they all won because they built an audience. The side hustle was the vehicle, but the audience was the fuel.” That’s an honest thesis. It’s also the one that undercuts the headline, because building that audience is slow, unpaid work that no chatbot shortcuts.

The proof on offer is a parade of individual peaks. Sean’s “$500,000 in a single month” on KDP. Josh going from under $1,000 a month to “$186,000 in a single month.” Carla’s single $106,000 contract from a 1,400-subscriber channel. And a one-person, AI-built telehealth company that supposedly did “$41 million in revenue” in year one. Hummus even pauses to note that founder “is under a ton of heat right now for cutting corners with AI, fake ads, and the whole mess.” That aside matters more than he lets on.

What the method actually requires

Start with the telehealth example, because it’s the most checkable. The company is Medvi, founded by Matthew Gallagher. Independent reporting puts its 2025 revenue at about $401 million — not the $41 million figure in the video — with a projection of $1.8 billion for 2026 on a skeleton staff, according to eMarketer. But the same reporting documents why it’s a cautionary tale, not a template: paid ads featuring fabricated doctors, an FDA warning letter in February 2026 over how it marketed compounded weight-loss drugs, and a class action. Hummus says “don’t do what he did.” Fair. Just notice that the single most spectacular number in the video is attached to a business now defined by regulatory trouble.

Now the tools, which the video gets mostly right. Claude’s $20-a-month Pro plan is real and is genuinely enough to draft a book, script videos, or write outreach emails. Claude is good at long-form writing. None of that is in dispute. The gap is between production and distribution — and the income lives entirely on the distribution side.

Take the “silent bookstore.” Amazon KDP pays either a 35% or a 70% royalty, and the 70% tier only applies to eBooks priced roughly $2.99–$9.99, with a per-megabyte delivery fee subtracted, per Amazon’s own pricing page. A $4.99 book at 70% nets you about $3.34 a sale. To clear $3,000 a month — less than many salaries — you need roughly 900 sales every month, forever, which means rank, reviews, and a reason for Amazon’s search to surface you over the flood of other AI-assisted books. Claude writes the book in a weekend. It does not generate those 900 monthly buyers.

The “audience machine” has a harder gate still. YouTube’s Partner Program won’t pay you ad revenue until you hit 1,000 subscribers and 4,000 valid public watch hours in 12 months, per YouTube’s official eligibility rules. Most channels never get there. The realtor whose video “hit over 101,000 views” is the survivor you’re shown; the far larger number of channels that posted consistently for a year and stalled at 200 subscribers don’t make the highlight reel. Hummus actually concedes this with the Carla story — most of her videos got “fewer than 200 views” — but frames the six-figure contract as proof the model works, not as the rare outcome it is.

Here’s the pattern across all five hustles:

The right-hand column is the job.

It’s unpaid for a long stretch, and it’s where almost everyone quits.

Who actually wins this game?

The winners in the video share a profile the pitch never spells out. Several already had domain expertise worth paying for — Josh in IT and cybersecurity, Isaiah as a singing coach, Nicole in compliance. Hummus himself has, by his own account, built businesses “to over $10 million” and has a community feeding him these case studies. That’s not a knock; it’s context. A person with a marketable skill, a few years of credibility, and the patience to post into the void for six months has a real shot here. A beginner with no skill to teach and no audience is starting from a much colder start than “turn your camera on and talk” suggests.

There’s also a selection effect baked into every example. You’re shown Sean’s $500,000 month, not the median KDP author, who earns very little. You’re shown the channel that took off, not the thousands of identical-sounding ones that didn’t. That’s survivorship bias doing the heavy lifting, and it’s worth naming plainly: the existence of a winner tells you the game is winnable, not that it’s likely.

What you’d realistically earn

Be honest with yourself about the curve. For the KDP and YouTube routes, a realistic first six months for someone starting cold is roughly $0 to a couple hundred dollars a month, because you’re below the thresholds where money even starts. A coaching or service business can pay faster — if you already have the skill and can land two or three clients — but “faster” still means weeks of outreach, not a same-week paycheck. The video’s framing (“pay more than a 9–5”) describes the ceiling for the rare operator who sticks it out for a year or more, not the floor for a beginner.

And the income-claim space is regulated for a reason. In the U.S., the FTC has been hammering “AI-powered passive income” pitches: it moved to shut down Click Profit, a scheme that cost consumers at least $14 million, and CNBC reported the agency’s case against an Amazon “passive income” operation that allegedly defrauded users of millions. To be clear, Hummus’s video isn’t one of those schemes — he’s selling a free training and a coaching offer, not a fake storefront. But the regulatory backdrop is the reason any “AI does it for you” income number deserves a calculator before it deserves your savings.

Who this is (and isn’t) for

This makes sense if you already have a skill or trade people pay for, can spare maybe 4–6 focused hours a week, and can stomach earning nothing while you build a body of content and a reputation. If you treat the YouTube channel or the books as a multi-month investment rather than a this-week payday, the model is sound and Claude genuinely cuts the production grind. It does not make sense if you’re hoping AI replaces the audience-building entirely, if you need income within 30 days, or if you have no underlying expertise to package. For that person, the realistic outcome is months of effort and a trickle of sales — and the temptation to buy a course to fix it.

What to remember

The video’s core claim survives, but only in a narrower form than the title promises: these five businesses can out-earn a salary, and Claude is a legitimate force multiplier for the writing inside them. What the pitch soft-pedals is that all five run on the same engine — a trusted audience — and that engine takes months of unpaid, un-automatable work to start. For more on the same creator’s framing, see our look at 10 Claude AI side hustles that can pay a full-time income and the reality behind selling ebooks with Claude. Claude can build the vehicle. You still have to find the road.