This is a short series on what it actually looks like to build software as a solo developer in 2026—with AI as a co-pilot, not a replacement.
I’m not here to sell you a toolchain or claim I’ve cracked some secret code. I’m just writing down what works, what doesn’t, and what I’m still figuring out.
Over the next few posts, I’ll cover:
- My current AI development workflow — the actual tools and habits I use day-to-day
- Taste vs. throughput — why AI can 10x your output and still make things worse
- Shipping before the system is perfect — the hardest lesson for an obsessive problem-solver
If you’re also building alone—or thinking about it—these notes are for you.
Why write this now
I’ve been shipping with AI assistance for over a year now. Not as a novelty. As the default mode of work.
It started with Vertex Platform Solutions, my solo dev studio. Then CareTender Intel, a procurement SaaS. Then Dev Foundry, a boilerplate to help others ship faster. Then a marathon training dashboard, a tmux voice dictation tool, an MCP server for Google Search Console, and more.
Each project taught me something different about where AI helps and where it gets in the way.
The pattern I’ve noticed: the builders who are thriving aren’t the ones with the best prompts. They’re the ones with the strongest taste—and the discipline to ship before they’ve optimised every edge case.
That’s what this series is about.
Who this is for
- Solo developers using AI to move faster
- Technical founders who feel like they’re drowning in tooling options
- Anyone who’s worried AI is making them a worse engineer, not a better one
- People who love building but struggle to call something “done”
The posts
- My AI Development Workflow — The concrete tools, habits, and mental models I use every day
- The Trap of AI-Assisted Mediocrity — Why more output isn’t always better output
- Shipping Before the System Is Perfect — a future post about learning to let go
Let’s get into it.