But again, it looks like we won't need as many developers in the near future. Instead, we'll need people who understand principles, architecture, and system design. We'll still need people with strong computer science backgrounds.

But the good news is that now even small teams with product ideas will be able to try create and test something really quickly.

It’s getting scary actually. Before that I was using AI just for productivity improvement and I just asked it to explain something or provide some examples, etc. I think we are just a couple of years away from the point AI will replace many developer.

Almost all of the code was written by AI agent (cline/claude) I just guided it through and asked what I want, but without going into details. The code is not perfect, and some may even say it is far from it. But it does the thing. And it took just a couple days to create the app.

I might also be hitting some bug in the macOS Bluetooth stack

I made some changes by switching the proximity detection algorithm from moving average to Kalman filtering. However, I'm still not completely satisfied with the results. I suppose 100% reliability isn't possible with Bluetooth.

One of the benefits of LemonSqueezy is that they are a MoR and they manage taxes for you. But Stripe has a very similar service (Stripe Tax), and I think they could integrate the rest of LemonSqueezy's features, like license and subscription management. But we'll see.

That's what I'm afraid of. I bet Stripe will eventually kill them. I was thinking of creating an intermediary license management backend, but it would increase the level of complexity a bit too much.

But the main thing is the fear of… failure

And it seems it’s a bit harder to prepare an app to sell outside of App Store. License management (the logic is even more complex than the core), automatic updates (not that hard actually, thanks to Sparkle).

At some point I started to think what to do next. I can outsource it, but instead I now want to try to sell it. It’s actually just another challenge - to not just create something, but sell it and make some profit. I decided to do that outside of App Store, mainly because it uses private APIs.

I quickly created some simple prototype without any GUI at first. I didn’t even try to look for already existing alternatives. It was some sort of a challenge. I was using it for quite a while, and it seems to be working quite well most of the time. Considering the nature of the Bluetooth.

And no os can do that for you automatically. macOS can unlock your Mac using a watch, but can’t lock it. That’s how I started to learn Swift and SwiftUI.