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

By the Income Reality Check Editorial Team · Updated 2026-11-12 · 7 min read

9 Claude AI night side hustles: the numbers are real, but who actually earns them?

Verdict: Misleading — the headline number is real but unrepresentative. Nearly every figure in the video checks out, yet almost all of them belong to top-1% outliers or to people whose income comes from YouTube, not the hustle named.

Shane Hummus opens “9 High Paying Claude AI Side Hustles You Can Start At Night” with a promise: a laptop, Claude, and the hours after everyone else goes to sleep are enough to build a second income without quitting your job. He backs it with real people and real dollar amounts flashed on screen. The striking part is that most of those numbers are accurate. The problem is who they belong to — and what the video quietly leaves out about how any of them were actually reached.

What the video actually claims

The pitch is nine “night” hustles, all powered by Claude: an AI romance-novel factory, teaching people how to get into your job, starting a YouTube channel, printable paper escape rooms, personalized children’s storybooks, Notion templates, Teachers Pay Teachers worksheets, building websites for local businesses, and finally “post about any of the above on social media.”

Each gets a flagship example. Romance author “Coral Hart” fed Claude her outlines, published more than 200 novels under 21 pen names in a year, sold roughly 50,000 copies, and cleared six figures — a story the New York Times really did run. Thomas Frank made over $1 million in a year selling Notion templates (Business Insider checked his documents). Kindergarten teacher Becky Powell made about $125,000 on Teachers Pay Teachers in 2022. James Pelton built local-business sites with Claude in about 45 minutes each and charges $49 a month. Hummus also stacks his own community members — Nurse Jen, his brother Zach, a product manager named Richard — who hit $6,000 to $12,000 months on YouTube while working day jobs.

Every number here is real, Hummus says, “and it’s sourced with real examples.” On that narrow point, he’s mostly right. The sleight of hand is in the framing.

What the method actually requires

Start with the romance factory, because it’s the cleanest case. Coral Hart’s six figures came from publishing over 200 books in a year across 21 pen names. Hummus tells you you’ll “finish about one book a month” — 12 a year. He’s quietly handed you a sixteenth of the output engine and kept the same income halo hovering above it. On Amazon KDP, a novel priced in the $2.99–$12.99 band earns the 70% royalty; below or above that range it drops to 35%. A $4.99 book nets you roughly $3.40 after the delivery fee. To clear even $2,000 in a month at that price, you need about 590 sales — of a brand-new pen name with no reviews, no ranking, and no readership, in the most saturated corner of self-publishing. Claude writes the draft in 45 minutes. It does nothing about discoverability, and discoverability is the whole game.

The YouTube examples carry a different hidden cost: the on-ramp. To earn a cent of ad revenue you must clear the YouTube Partner Program threshold — 1,000 subscribers and 4,000 qualified public watch hours in 12 months (or 10 million Shorts views in 90 days). That’s the floor, not the goal. And note what’s buried in the “two hours a month” story about Richard: these are coaching clients of Hummus’s program. Their income is YouTube AdSense plus courses plus consulting — not the underlying worksheet or escape room. The video uses “side hustle” success stories whose actual revenue engine is the thing Hummus sells at the end: help growing a YouTube channel.

Look at how skewed the platforms really are. The table below pulls median versus top-tier earnings for the three “sell-a-file” hustles the video leans on hardest.

Thomas Frank’s own account, which Business Insider verified against his revenue documents, is the tell. His million-dollar year didn’t start with a template. It started with nine years of YouTube and 2.8 million subscribers — a distribution asset built before the first template went on sale. Strip the audience out and you have a $49 file nobody can find.

Who actually wins this game?

The winners cluster into three groups, and the video’s own examples prove it. First, people who already have an audience — Thomas Frank is the textbook case. Second, early movers and high-volume operators: Coral Hart’s edge isn’t one book a month, it’s an industrialized 200-book pipeline across 21 brands, run before AI fiction flooded Amazon (one tracker counted roughly 1,100 AI novels hitting the store in a single week this year). Third, people with real domain expertise they’re repackaging — Nurse Jen spent 33 years explaining U.S. healthcare to patients, Zach spent 30 years in the trades. Claude didn’t give them the knowledge or the audience. It compressed the production step, which was never the bottleneck.

Notice who’s missing from that list: the beginner with no audience, no niche authority, and a sixteenth of the output. That’s the person the thumbnail is aimed at.

What you’d realistically earn

Here’s the honest range the video won’t give you. On Teachers Pay Teachers, most sellers earn under $500 a year; committed sellers who treat it as a near-full-time side venture report maybe $300–$1,000 a month after a year of consistent work, and the $100k+ tier is roughly the top 1%. Notion template sellers mostly make a handful of sales unless they have traffic. A fresh KDP pen name typically sells in the single digits per title until something ranks.

For most people starting from zero, a realistic first six to twelve months looks like $0–$200 a month, with a plausible climb to a few hundred — occasionally low four figures — after a year of steady nights and, crucially, after solving distribution. That shape matches what practitioners told CNBC: “passive” income usually means a long stretch of unpaid upfront work before any daily workload drops. Compare that to the on-screen anchors of $125,000, $1 million, and “a $78,000-a-year run rate in 29 days.” The gap isn’t a lie. It’s the difference between a documented outlier and a typical outcome — exactly the distinction the FTC says matters most. U.S. regulators treat earnings claims as deceptive when they parade atypical results as representative and don’t account for what participants actually spend and typically make. The FTC’s 2025 proposed Earnings Claim Rule would extend that scrutiny to business-coaching and “money-making opportunity” marketing specifically.

To his credit, Hummus almost says the quiet part out loud near the end: “just about every side hustle works, but they all have one thing in common — they need traffic.” That single sentence is the real content of the video. Everything before it is nine ways to make a product. Only the last 90 seconds admits that making the product was never the hard part.

Who this is (and isn’t) for

This set of hustles makes sense if you already have one of three things: an audience to sell to, deep expertise worth repackaging, or the patience and budget to build distribution over a year — think 8–15 hours a week for 12 months before meaningful money, plus either ad spend or a relentless social-posting habit. If you enjoy writing romance, teaching, or making templates for their own sake, the downside is mostly your time. It’s a poor fit if you need income this quarter, have no existing audience, and are picturing the on-screen numbers as a near-term target. The “at night, no skills needed” framing is where the expectation and the reality separate. For a fuller breakdown of how these Claude pitches tend to work in practice, see our look at 10 Claude AI side hustles that can pay a full-time income and the specific KDP angle in a “once in a lifetime” KDP opportunity.

What to remember

The figures in this video are, unusually, verifiable — and that’s what makes it misleading rather than false. Shane Hummus shows you real top-1% results, hands you a fraction of the machine that produced them, and leaves out the distribution work that every one of his own examples quietly depended on. Claude genuinely collapses the production step from a weekend to an hour. The selling, ranking, and audience-building it doesn’t touch is still the entire job.