The real story isn't that normies are building apps with AI — it's that they're solving problems too niche for any startup to care about.
The Summary
- Non-coders are using AI to build custom apps for hyper-specific personal problems, from wedding seating charts to Pilates trackers — what they're calling "vibe coding"
- A third of Americans now interact with AI weekly, but most use cases still feel like overkill for basic questions
- The interesting signal: people are building practical tools for problems that have no market, no VC interest, and no off-the-shelf solution
The Signal
Shayan Mirzazadeh failed computer science twice in college. Now he's an account manager who builds apps with AI for work pain points and personal projects, including a Pilates flow tracker for his fiancée. His co-worker Jayne Ingram-Roberts built a fantasy-league-style app for Big Brother fans. Their biggest project is Seatbee, a wedding seating chart generator that enforces user-defined rules about who sits where.
Over 200 people are already using Seatbee. That's not a unicorn. That's not even a viable startup. But it's exactly the kind of problem that has no other solution because no company would ever build it.
"Wedding seating charts are too niche for SaaS, too personal for templates, and too annoying to do by hand."
This is the first wave of what actually matters about accessible AI tooling. Not chatbots answering questions you could Google. Not writing your emails in a slightly different tone. The value is in the long tail of human problems that never had enough economic gravity to attract a developer.
Key shifts happening:
- Market failures become buildable: Problems affecting 200 people can now have custom software solutions
- Personal automation at scale: Every household annoyance is potentially solvable if you can describe it
- Zero to functional in hours: No bootcamp, no GitHub, no Stack Overflow rabbit holes
The comparison to coding custom Tumblr themes is spot-on. That era trained millions of people to tinker with code for purely personal aesthetic reasons. Vibe coding does the same thing, but for functionality instead of decoration. You're not learning to be a developer. You're learning to be specific about what you want.
The economics here are worth watching. When the cost of building software approaches zero, what happens to the software industry? Not the AI infrastructure layer, which prints money. The application layer. The SaaS companies charging $29/month for tools that solve problems 80% of the way.
Why this matters now:
- AI coding assistants are good enough that intent matters more than syntax
- The barrier isn't technical skill anymore, it's knowing what you want built
- Every person with a recurring annoyance is now a potential builder
The skepticism about AI use cases has been justified. Most of what people do with ChatGPT is intellectually lazy or marginally useful. But vibe coding occupies different territory. It's not replacing thinking. It's replacing the gatekeeping that kept non-technical people from automating their own lives.
The Implication
Watch what gets built in the margins. The wedding seating app today. The hyper-specific inventory tracker tomorrow. The custom workflow tool that saves someone 30 minutes every week but would never justify hiring a developer.
This is how AI gets embedded in daily life — not through one killer app, but through a thousand weird little tools that didn't exist because they weren't worth building. Until now. If you've got a recurring problem that's too small to pay someone to fix but big enough to annoy you weekly, you now have options. The question is whether you're willing to learn enough to describe what you want clearly.