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

By the Income Reality Check Editorial Team · Updated 2026-09-19 · 7 min read

Claude AI side hustles that “pay a full-time income”: the receipts Shane Hummus skips

Verdict: Half-true — works only if you do the unspoken work. The five jobs are real and Claude genuinely speeds them up, but every income number is a peak performer, and the video’s real product is a YouTube coaching program.

Shane Hummus opens “5 Boring But Easy Claude AI Side Hustles That Can Pay a Full-time Income” with a promise: no degree, no experience, no startup money, and “a few of them pay more than the average doctor’s salary.” He’s pitching genealogy research, selling AI prompts, AI bookkeeping, Excel consulting, and thumbnail design — all run through Claude. Are these real ways to make money? Yes. Will they pay a full-time income the way the video implies? That depends almost entirely on the work he glosses over in the phrase “let’s talk about how to do this with Claude.”

What the video actually claims

Each hustle comes wrapped in a specific, name-attached number. Genealogists bill “$30 an hour to well over $200 an hour,” and a firm called Legacy Tree “bills from $108 an hour in packages of 25 to 100 hours.” Prompt sellers keep 80% on PromptBase, and Hummus points to Anthropic advertising a prompt-engineer role “paying up to $335,000 a year” plus a CNBC-covered art-history PhD earning over $100,000. A YouTube bookkeeper he calls Pascarelloo supposedly manages 87 clients at “$547 per client recurring,” about $47,000 a month. An Excel freelancer named Chandu “made $100,000.” Mr. Beast reportedly pays “$10,000 per thumbnail.”

The Claude angle is the same each time: open a project, upload the messy inputs, let Claude transcribe, model, draft, or brief, then you “sanity check” and invoice. His honest through-line — and it’s a good one — is “you bring the hustle, Claude brings the skills.” AI handles production; a human still has to find clients and verify output.

Here’s the part the receipts don’t advertise. Woven between the five hustles are repeated pitches for a free “live training” this week, a “niche validator” giveaway, and an application to “one-on-one mentorship or literally done-for-you work” that accepts “about 18% of people who apply.” The income stories that carry the most emotional weight — Nurse Jen’s $10,500 month, his brother Zach’s $18,000 month — aren’t about the five Claude hustles at all. They’re about starting a YouTube channel through his program.

What the method actually requires

Take the numbers one at a time and they hold up as ceilings, not averages. The Association of Professional Genealogists does say rates run from around $30 to well over $200 an hour, and notes no license exists (APG). But that same range means many working researchers sit near the bottom, and $200/hour reflects credentialed specialists with published casework — the “finished research is the credential” cuts both ways, because you build that portfolio unpaid or underpaid first.

Bookkeeping is where the gap between claim and norm is widest. Hummus cites one YouTuber pulling roughly $47,000 a month from 87 clients. The U.S. Bureau of Labor Statistics puts the median wage for bookkeeping, accounting, and auditing clerks at $50,670 a year — about $24.36 an hour — and projects the occupation to decline 6% between 2025 and 2035 as software automates routine entry (BLS). So the same automation Hummus sells as your edge is the reason the field is contracting. Independent freelance bookkeepers typically build a roster of 10–15 recurring clients at $200–$800 a month each, and that roster takes months of outreach to assemble. Eighty-seven clients isn’t a starting point. It’s a small firm with staff.

The prompt-engineering figure is the cleanest bait-and-switch, though probably an unintentional one. Anthropic’s $335,000 posting was a full-time salaried job in San Francisco — a W-2 role with interviews and coworkers — not money you earn selling $5 sentences from your kitchen. Independent trackers put typical prompt-engineer pay closer to $125,000, and that’s still for employment, not a marketplace side hustle. PromptBase sellers do keep 80%, but on prompts that often list for a few dollars; clearing real money there means volume and marketing the video never quantifies.

Who actually wins this game?

Look at who succeeds in every one of these lanes and a pattern shows up fast. It’s people who already have the domain expertise Claude can’t supply. Hummus even admits this in his best example: he cut his realtor sister-in-law’s paperwork from two hours to ten minutes, “but the only way I was able to help her do that is because of the fact that she was an expert.” The prompt, the Excel model, the thumbnail brief — Claude executes them, but the value comes from knowing what “correct” looks like. A finance background makes the Excel gig pay. Real estate knowledge makes the prompt sellable. Taste, not rendering, is what earns $1,000 for a thumbnail.

The second group that wins is people who can market. Every income story in the video is really a marketing story — the bookkeeper films his work, the QuickBooks tutorial earns AdSense on top of the bookkeeping, Nurse Jen’s money comes from a YouTube channel. The hustle isn’t the bottleneck. Distribution is.

What you’d realistically earn

If you already have a marketable skill and you commit to client outreach, these are legitimate freelance paths — genealogy, bookkeeping, and Excel consulting have existed for decades and Claude makes each one faster. A realistic first year for a beginner starting from zero looks like a few hundred dollars a month while you land your first two or three recurring clients, climbing toward $1,000–$3,000 a month after a year of consistent selling. That’s a real second income. It is not “more than the average doctor’s salary,” and almost none of it is passive in year one.

The full-time-income framing leans on outliers. One bookkeeper at 87 clients, one Excel guy at $100,000, Mr. Beast at $10,000 a thumbnail — these are the visible tip, not the median. U.S. readers should also know the regulatory backdrop: the FTC has proposed a new Earnings Claims Rule and an expanded “Business and Money-Making Opportunity Rule” (16 CFR 437) specifically because “phony claims about likely earnings lure people… into spending thousands” on business coaching (FTC). That’s not an accusation against this creator — it’s the standard any income pitch tied to a paid program should be measured against.

Who this is (and isn’t) for

This works for you if you bring an existing skill (accounting, a trade, real estate, design taste), have 8–15 hours a week for unglamorous client hunting, and treat Claude as a speed multiplier on work you could already sanity-check yourself. It works poorly if you’re expecting the “no experience needed” line to be literally true, if you have no time to market, or if you’re drawn mainly by the doctor-salary hook. And note the YouTube monetization stories: joining the Partner Program requires 1,000 subscribers plus 4,000 watch hours in a year — rising to 8,000 hours in February 2027 (YouTube Help) — which is months of unpaid uploading before the first AdSense dollar.

What to remember

The five hustles are real, the demand is real, and Claude honestly does compress the boring parts. What’s missing is the distance between a peak-performer receipt and a median outcome, and the fact that the video’s warmest success stories are selling a separate YouTube program, not the Claude workflows themselves. Bring a skill, expect to market it for months, and treat every named dollar figure as a ceiling. For more on how these pitches are built, see our looks at 10 Claude AI side hustles that “pay a full-time income” and 8 Claude AI side hustles that “replace a full-time job”.