My AI-Assisted Coding Journey

The Starting Point: HTML from 2003 and Big Dreams

I'm a an ordinary family man from Norway with what I'd call a "normal tech-related job". Not a programmer, not a developer—just someone who remembers making terrible websites with HTML and CSS when that was still exciting around 2003.

Fast forward to late 2022: ChatGPT and Midjourney dropped, and like everyone else, I was blown away. I played around like most people did—asking it to write emails, generate ideas, maybe help with some work tasks. Fun, but nothing life-changing.

Spring 2025: The Prompt Optimization Rabbit Hole

Then something clicked this spring. I got genuinely interested in prompt optimization—not just using AI, but getting really good at using AI. Suddenly I could prototype simple MVPs in React, build decent HTML prototypes, get real help with actual problems.

It wasn't magic. It was systematic. I built up a library of techniques, learned which models were good for what, figured out how to get AI to teach me basic developer skills.

The Great App Brainstorm: 30+ Ideas

Armed with these new skills, I went completely overboard. I brainstormed and prototyped over 30 app ideas. Some were terrible, some close to pretty good, most were somewhere in between.

But there was this recurring theme: constraint as liberation. Apps that did less, but did it better. Tools that served instead of captured attention.

Five ideas kept coming back:

• FocusAnchor - One task at a time, present moment focus
• NoteToSelf - Maximum 11 notes total (1 active + 10 archived)
• Mikro - Life logging that tracks patterns, not metrics
• MemoryAnchor - Spatial memory organization with dual modes
• DayRater - A simple app for daily reflections

The Summer Experiment: Could I Actually Do This?

Summer vacation 2025. The family was heading into summer mode, and I had a wild idea: What if I spent the summer seeing if I could get one app "App Store ready" using AI-assisted coding?

Not "learn to code" and then build an app. Use AI as my development partner from day one.

"If I get stuck with one AI model, I'll try another. If ChatGPT can't solve the SwiftUI layout issue, maybe Claude can. If neither works, maybe Cursor can."

What Actually Happened: The Wildest Summer Ever

Here's where the story gets crazy.

I didn't just get one app ready. I ended up with:

• 5 apps live in the App Store (FocusAnchor, NoteToSelf, Mikro, MemoryAnchor, M43 DoF)
• DayRater in App Store review (the 6th app!)
• A complete brand (DigTek - "Less, but better")
• This website you're reading right now

Each app embodies the same philosophy: constraint as liberation. One powerful feature done beautifully, instead of 50 features done poorly.

The AI Development Process That Actually Worked

Here's what I learned about AI-assisted app development:

• Start with constraints, not features - "What's the ONE thing this app does?" is more powerful than "What could this app do?"
• Use different AIs for different strengths - Claude for SwiftUI architecture, ChatGPT for debugging, Cursor for rapid iteration
• Build systematically - Create reusable components, establish patterns, document what works
• Embrace being a beginner - Ask dumb questions. AI doesn't judge you for not knowing basic concepts
• Focus on user experience over technical sophistication - Better to have a simple app that works than a complex app that crashes

The Norwegian Reality Check

Let me be clear: this isn't a "quit your job and become an app developer" story. I'm still a family man with a normal job. These aren't going to replace anyone's salary anytime soon.

But there's something profound about taking an idea from "wouldn't it be cool if..." to "here's the App Store link." AI made that possible for someone like me.

What This Means for Everyone Else

If you have an idea for a simple, focused app—and you're willing to learn prompt engineering rather than traditional programming—this is probably the best time in history to try building it.

You don't need a computer science degree. You need:

• A clear vision of what problem you're solving
• The discipline to say no to feature creep
• Patience with the learning curve (it's still work!)
• Access to good AI models and the willingness to use multiple ones

The Apps Are Live. Now What?

FocusAnchor, NoteToSelf, Mikro, MemoryAnchor, and M43 DoF are all live in the App Store. DayRater should be approved soon. Each one embodies our "constraint as liberation" philosophy.

Will they change the world? Probably not. But they might change how a few people think about productivity, memory, and focus. And honestly, that's enough.

This summer taught me that the biggest constraint isn't technical knowledge—it's the courage to start with constraints instead of features.