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

By the Income Reality Check Editorial Team · Updated 2026-07-15 · 7 min read

Claude Code built a $20K membership site — but who was already watching?

Verdict: Misleading — the headline number is real but unrepresentative. The AI tool did the building. A decade-old audience did the selling.

In a conversation on The Calum Johnson Show, designer Oliur (Oliur Online) explains how he used Claude Code — Anthropic’s AI coding tool — to build a membership site called digitalcreatorclub.com “in like a weekend,” and how it did “like $20,000” in its first week. He says he has no background in databases or APIs. He says anyone can do it. The build part is true. The reason it made $20,000 is the part the title leaves out, and it’s not the AI.

What the video actually claims

The pitch has two layers. The first is about the tool: Claude Code, used through the Claude desktop app rather than the intimidating terminal, can turn plain-English prompts into working software. Oliur demonstrates it live, spinning up a to-do app and then restyling it with an “Apple aesthetic” in a few minutes. That demo is honest. The tool really does that.

The second layer is the money. Oliur says the membership platform — payments, a database, invite logic, Discord onboarding, member content — would have cost “$50,000, $100,000” and a hired engineering team before, but he built it himself over a weekend and it “made like $20,000” in week one. He later estimates the club has earned “$40 to $50,000” total, “with very minimal advertising.” He also mentions a $30,000 month selling Tumblr templates around 2013, and $30,000 from AI-generated wallpapers.

To his credit, Oliur doesn’t oversell the automation. He admits things break, that he screenshots errors and pastes them back in, and that “there’s no such thing as one-shotting anything.” The problem isn’t dishonesty. It’s framing. A viewer walks away thinking the tool produced the income. It didn’t.

What the method actually requires

Start with the tool cost, because it’s the easy part. Claude Code isn’t a separate purchase — it’s bundled into a Claude subscription. Anthropic’s pricing page lists Pro at $20/month (or $17 billed annually) and Max from $100/month, with Claude Code included in both. So the software that “would have cost $100,000” costs $20 a month. That claim is fair, and it’s the strongest part of the video.

Now the hard part: distribution. A membership site is a subscription business, and subscription businesses live and die on two numbers the video never mentions — how many visitors convert, and how fast members leave. Industry benchmarks put visitor-to-member conversion around 2%, and monthly churn for content and creator communities in the 3–6% range. Do the arithmetic. To clear $20,000 in a week at, say, a $99 lifetime or annual price, you need roughly 200 buyers. At a 2% conversion rate, that’s about 10,000 motivated visitors — in seven days.

Where does a brand-new site get 10,000 buyers in a week? It doesn’t. It gets them from an audience that already exists.

That’s the missing input. Oliur says the club idea “came from my own audience” who’d been asking for it “for a while now.” He mentions a “tech YouTube channel” with brand sponsorships, a blog with posts going back to 2015, and “nine different products, services, newsletters or affiliates” housed on his personal site. The $20,000 didn’t come from Claude Code. It came from a launch email — or a video, or a post — to people who already knew and trusted him. Swap in an unknown founder with the identical code and the first-week number is closer to zero.

CNBC has made this point repeatedly with actual operators. A creator earning $160,000 a month in “passive” income told the outlet that one of the “complete lies” about side hustles is that they run on autopilot; another said each piece of content takes about two hours and that success followed more than 1,000 videos (CNBC). NerdWallet’s guidance is blunter still: the durable income streams aren’t passive, they’re “strategically automated businesses” built on upfront and ongoing work (NerdWallet).

Who actually wins this game?

The people who win with Claude Code the way Oliur did share one trait, and it isn’t coding skill. It’s an audience they built before the tool existed.

Oliur’s own timeline gives it away. The Tumblr templates that made $30,000 in a month? That was around 2013 — early-mover timing on a platform before it was saturated. The wallpapers, the YouTube channel, the blog, the newsletter — a decade of accumulated attention. He even frames a personal website as “internet points” and a “public CV,” which is exactly right and exactly the point: he already had the points. AI collapsed his build time from weeks to a weekend. It did nothing for his distribution time, because that was already paid off, years ago.

So the honest winners here are (a) established creators with a list to launch to, (b) operators willing to buy traffic, and (c) people with real domain expertise who can make a product worth paying for. A beginner with a Claude subscription and no audience has the cheapest input solved and the expensive one — attention — completely unsolved.

What you’d realistically earn

Let’s be concrete about the gap between the headline and a beginner’s likely reality.

If you’re starting with no audience, expect months of $0 to a few hundred dollars while you build the thing that actually drives sales — the email list, the search rankings, the social following, the reputation. The creator-economy data is sobering: across an estimated 27 million paid U.S. creators, more than half earn under $15,000 a year, only about 12% of full-time creators clear $50,000, and roughly 4% reach $100,000. A first-week $20,000 launch sits far out in the tail of that distribution, and it’s powered by pre-existing reach, not by the AI build step you can replicate for $20.

None of that makes the tool worthless. Building a membership site, a personal site, or a small app for $20/month instead of $50,000 is a genuine shift, and it lowers the cost of trying. Just don’t confuse a lower build cost with guaranteed revenue. The build was never the bottleneck for most people. Getting anyone to show up is.

One more thing worth flagging for U.S. readers specifically: the FTC is actively tightening the rules on “make money” content. In January 2025 the agency proposed a new Earnings Claim Rule and expanded its Business Opportunity Rule to cover “money-making opportunities,” requiring sellers to substantiate income claims with written proof available on request (FTC). This particular video is a case study, not a product being sold, so it’s a different category — but if you ever put “I made $20K in 7 days” on something you’re charging for, that number now needs receipts.

Who this is (and isn’t) for

This makes sense for you if you already have some audience — even a small newsletter or an engaged following — plus a few hours a week and a specific thing people have asked you to make. In that case Claude Code genuinely removes the build barrier, and the $20 subscription is one of the better bets in this whole category. It also makes sense as a skill-building exercise: build the personal website Oliur recommends, learn how the tool thinks, and lower your own cost of experimenting.

It’s a poor fit if you’re expecting the tool to supply customers. If you have no audience, no traffic budget, and no niche expertise, you’ll build something functional in a weekend and then watch it sit there, because nobody knows it exists. The build was the 10% Claude solved. The 90% is still yours.

What to remember

Oliur’s $20,000 week is almost certainly real, and Claude Code almost certainly built the site. Those two facts are true and unrelated in the way the title implies. The income traces back to a decade of audience-building the video treats as a footnote, and the tool that gets top billing is the cheap, easy, already-solved part of the equation. Real capability, real number — attached to the wrong cause. If you want to build with it, great. Just budget your time for the part no AI does for you: getting people to care.

For more on where these tools actually help versus where the marketing gets ahead of reality, see 10 Claude AI side hustles that can pay a full-time income and how I’d make money with Claude AI in 2026 if I had to start over.