Codex Skills for iOS Development · Replay network requests inside Swift Tests · LLM improvements 12 months later

It almost feels embarrassing to admit, but I've started another app project. If you asked me a year ago what I'd recommend doing for success, one of the things would be to focus on a single thing.

More than ever, I feel like I'm reinventing myself. I still believe in focus, and I'm still consciously deciding where to spend my time. However, I also find myself developing differently.

I'm embracing AI and I'm doing a lot of development with agents these days. I'm in the process of optimizing my workflow, optimizing it for higher code quality delivery (and yes, I'll share you all about it once I feel like I have a solid one). This also means that I find myself waiting for AI results more and more often. It's not exceptional for Cursor to run minutes before it completes on a large refactor or feature development.

If you know me, you know I want to make the most of my time. Instead of scrolling through socials while waiting for Cursor to complete, I figured it would make more sense to trigger another AI run for another project. At first, and still, I would do this for RocketSim. Yet, spending time on both RocketSim and my 10K MRR idea was not enough to keep me busy.

What's interesting is that historically, I would write down app ideas for later revisitting. I simply didn't have the time to work on them, and I wanted to guard my focus for RocketSim & SwiftLee. Nowadays, however, it's so easy to get a proof of concept done by AI. For that reason, I'm also optimizing my project setup for reusability, so I can really quickly get going with a new project. This means:

Copying codex skills Copying Cursor rules Copying AGENTS.md files Copying linter rules

Basically all the things that help me get a solid AI output.

This new reality means I no longer have to schedule an app idea for later revisit: I can simply build it, try it, and evaluate how it feels.

So yes, I've started another app idea. It's early days, and I don't want to share much about it yet. I'm planning on dropping it in one of the future YouTube episodes when it turns out to become an app idea I truly want to release. I mean, nobody said I would build only one app idea towards 10K MRR, right?