Where Are All the AI-Built Apps?

A lot of people claim it's becoming simple to build websites and apps. Yet, I haven't seen a flood of new apps.

Part of the issue is that AI tools overpromise and underdeliver. Replit has made it simple to build prototypes for people who, only a year ago, couldn't have built that same prototype in months. But the apps it churns out are basic, often leaning on recycled or half-assed UX patterns that don't solve problems better than an existing free alternative. The future is bright for these tools, for sure, but they just aren't there yet for most people.

If they did deliver what they promise, why are we still using Facebook (if you're old), Twitter (if you're old), Instagram (if you're old), and whatever the kids use? Because those apps are very hard to build and AI is not that good. Yet.

A lot of folks think AI is just hype because they don't see a dramatic overnight change in platforms like Twitter or Amazon. But AI is making people faster and better, even if that change is hard to measure. Those who go beyond copycat products are launching apps to automate workflows, build internal tools, or serve small groups. The small scopes are easier to leverage AI on, and that's the point.

What we're seeing is the democratization of creating many small apps that are roughly the same as those that already exist—the to-do list makers, recipe lists, habit trackers, etc. This is partly because the AI can do these well because there are so many examples and because people are, rightly, so damn excited to be about to eject things from their brains onto a screen that they don't stop to care if a better version is a click away.

The unnoticed part is the people making simple, sometimes novel apps for small groups without caring about monetization. Tools like Bolt, Cursor, Replit—and even Claude Artifacts or ChatGPT Canvas—make it so ridiculously easy to create new apps. People spin them up, use them in tight groups, and don't worry about promoting them or trying to make money from them. This is quite a bit better than a group trying to cash in on the same stupid to-do list app.

At Mostly Serious, we play office bingo (actually called Bingoringo for reasons). We had a little app we used for years to run the game. An intern built it, and it was fine, but it lacked a handful of features I always wanted but we never got around to building. In a few hours, I built a new version with the functionality I wanted. We use it for our office bingo games, but if you don't work at Mostly Serious, you wouldn't know.

We also have an annual competition called The House Cup. This year, the theme is Star Wars and Spence built a great framework around it. In an hour or so over coffee, I spun up a site with a leaderboard to track our team's progress while bringing a little more life to it.

Last year, my friend Tom Douglas invited me to join a group that aimed to do 100 pushups every day. It seemed like a fun challenge, so I jumped in. A year and a ton of pushups later, it's become a habit. The problem is my memory is bullshit and I often forgot my count throughout the day, so I made a simple little app to keep track of my pushups. Eventually, I added a leaderboard and invited some friends. I'm about to wrap v2 that adds quite a bit more functionality, and I'll actually release it on the app store for other people who want to join in the fun. There's a donation link to help me cover hard costs and the amount of time I told Cursor to PLEASE STOP SUCKING SO MUCH AND JUST ADD THE FEATURE LIKE I'M TELLING YOU TO (still learning to become a leet prompt engineer).

But it's not about the money. There's a personal side to it all. Research says we can only keep a small circle of close friends, so I'd rather share Bingoringo or the House Cup with a handful of close friends and colleagues than blast it online.

And It's good to make things just because they deserve to exist. Not every new idea needs to become a unicorn startup. The world will be a better place if we bring social networks down to friend-sized AI-built apps.