The Setup: When a Tool and a Fantasy Collide
Last summer, Anthropic rolled out Claude Projects, and if you spend any time in the AI productivity space, you probably heard about it the way you hear about a new restaurant opening downtown: with cautious optimism and the sense that maybe this time things would be different. The feature promised something genuinely useful—persistent context, memory across conversations, the ability to feed an AI system a bunch of information and have it actually remember what you told it three weeks ago. It sounded like the scaffolding for something real.
I watched the launch unfold while simultaneously watching YouTube’s algorithm serve me content about AI side hustles. Thousands of videos. Billions of views, actually—over 1.2 billion last year according to tracking data. The passive income tutorials were the crown jewels of this content ecosystem, each one more confident than the last. Build digital products. Sell them on Etsy. Let the AI do the heavy lifting. Watch your bank account grow while you sleep.
The logic felt airtight to me, which should have been my first red flag. I had Claude Projects. I had Canva. I had an Etsy shop I’d created on a whim two years prior that had exactly zero sales. What I didn’t have was a realistic assessment of what I was walking into, but that’s never stopped anyone before.
The Project: Why It Looked Promising on Day One
I decided to create digital product templates. Specifically, social media content calendars, budget trackers, and habit-tracking dashboards—the kind of thing that looks good as a PDF and costs nothing to ship. Claude Projects let me set up a persistent workspace where I could dump design briefs, brand guidelines, and template specifications. I could ask Claude to iterate on designs, generate descriptions for Etsy listings, and brainstorm marketing angles. The AI would remember everything.
This is where I need to be honest about something: the tool actually worked. Claude’s Projects feature did exactly what it promised. I could maintain complex context across dozens of conversations. The AI generated thoughtful product descriptions. It helped me think through pricing strategies and came up with genuinely clever keywords for discoverability. The friction that usually exists between idea and execution? Reduced. That part was legitimately impressive.
So I did what the tutorials suggested. I created twelve templates over eight weeks. I designed them in Canva, paid for the premium tier to unlock the good fonts. I wrote product descriptions, set up SEO tags, and priced everything competitively. Underpriced them, actually, which in retrospect was the only thing I got exactly right. I was following the script so closely I could have been reading it aloud.
The Results: Where the Numbers Stop Being Theoretical
Three months of actual work resulted in four purchases. Three of them were from the same person who bought one template and then came back for the others. The fourth was from someone I didn’t know. Total revenue: twenty-three dollars. Total time investment: forty-seven hours. Do the math yourself if you want, but I’ll save you the effort. Forty-nine cents per hour.
There’s a meaningful question hiding in that number. When Influencer Marketing Hub’s AI side hustle report 2025 found that 71 percent of people attempting AI-assisted digital product creation earned less than a hundred dollars from their effort, they were looking at the tail end of a long distribution. I was just further out than most.
But here’s what stopped bothering me about my personal failure and started bothering me about the broader ecosystem: Etsy’s Q3 2025 earnings showed that AI-generated digital downloads had become the most-uploaded product category on the platform. They’d also become the lowest-value category—an average selling price of $3.40, down from $7.20 just two years earlier. The market had flooded. Massively. And Etsy’s Q3 2025 investor earnings release made it clear this was a trend, not a blip.
I wasn’t unlucky. I wasn’t missing some crucial insight. I was operating in a category that had become simultaneously saturated and devalued, using a tool that had lowered the barrier to entry so far that anyone with an idea and two hours could produce inventory. The AI didn’t create an advantage. It eliminated it.
The Uncomfortable Truth: Why This Matters
Let me say something that won’t be popular: Claude Projects is excellent. The tool works exactly as intended. The problem isn’t Claude. The problem is what happens when you give millions of people an excellent tool and tell them it’s their ticket to passive income. Everyone becomes a competitor overnight. Competition drives prices down. Lower prices mean the only way to make meaningful money is through volume, and volume requires marketing skills, audience-building, or luck. Preferably all three.
The AI didn’t fail me. The narrative failed me. The content ecosystem that profited from my optimism—the YouTube videos with the eye-catching thumbnails, the email newsletters about side hustles, the Twitter threads about building six figures—that ecosystem had already succeeded before I even started. They made their money from my attention and my hope. The passive income was theirs, not mine.
What gets me is that this isn’t some cynical take meant to trash the concept. Some people will build real businesses using Claude, using Etsy, using AI tools. Some people will actually hit five hundred dollars a month or way more. But they won’t do it by making one tool and hoping. They’ll do it through distribution channels that already exist, or by building an audience first, or by solving problems that people are actively looking to pay for. The shortcuts you see in the tutorials aren’t shortcuts. They’re just the middle part of the story, edited to remove everything before the success moment and everything after it.
What I Actually Learned (Beyond the Obvious)
I spent forty-seven hours and made twenty-three dollars, which is approximately forty-six hours and twenty-three dollars worse than I hoped. But I noticed something while doing this work that I wouldn’t have noticed otherwise. The best part of the process wasn’t the potential for money. It was the problem-solving. Figuring out how to structure a template so it was genuinely useful. Thinking about what features a budget tracker actually needs. Imagining the person who might use these things and whether I’d built something they’d actually want to keep.
That’s worth something, even if the Etsy shop didn’t validate it. The skill is real. The AI literacy is real. The ability to move from idea to product is real. I just learned it’s not worth $500 a month. It’s worth what it’s worth, which in my case was a clearer understanding of why passive income feels so appealing and why it rarely works the way we imagine.
If you’ve tried something like this yourself, I want to know what happened. Not the success stories—I can find those on YouTube. I want to hear about the projects that didn’t land, the time you spent, the moment you decided whether to keep going or cut your losses. That’s where the actual information lives.