On building
Ideas just got way more valuable
For most of my career, the thing holding me back wasn't ideas. It was access.
I've spent fifteen-plus years in operations. Warehouses, manufacturing, security, tech, government, my own company. Underneath, it's always the same job: figure out what's broken and fix it, figure out what needs created and go make it happen. Finding problems and seeing what was missing always came easy to me. Getting to a resolution or creating a system meant getting in line for a tech team, making the case, and waiting weeks to see if the budget or priorities supported it.
Then in 2025 I sat down with my oldest daughter to build a creative world with AI. Partly to show her this AI thing was going to matter for her. Partly to find out for myself what the tools could really do, instead of taking the hype's word for it.
The answer was: a lot more than I figured. I found out that the things that used to cost a budget and somebody else's quarter are now just my time and how fast I'm willing to learn. That changed the whole game for me.
Learnings
Ten months in: what building with AI actually taught me
Ten months ago I had never written a line of software. Since then I've built a handful of working things: a crochet charting tool, an AI system that runs whole projects on its own, and a couple of creative pipelines. Here's what that actually taught me, with the hype left out.
Coding is actually more about knowing what's worth making than the book side of knowing how to code something. AI knows how to build code pretty well now, and it will cheerfully build you the wrong thing and make it look great. Deciding what's worth building is still on you.
Finished and actually working long term is the only thing that counts. A demo proves something is possible. Getting it actually working, where a real person can use it without me standing there, is a different job, and that's where most of the work hides. AI loves to “get it out the door.” If you aren't careful you will make something that looks great but is a lie.
The tools change every week. What I knew in the spring was half-stale by summer. Keeping up isn't a thing you finish, it's just the work now. Some will find this frustrating, but if you love to learn, AI will not disappoint. I started an IBM course on AI and about a quarter of the way through saw that it was teaching concepts like RAG when everything had already moved up a layer. Agents, huge context windows, hosted abstractions that handle retrieval for you. Being taught to hand-wire what the stack already absorbed.
Most of what used to block someone like me wasn't technical. It was money and permission. Both are mostly gone. And if they're gone for me, a non-technical regular guy, they're gone for a lot of people who haven't noticed yet.
That last part is really why I started writing any of this down. I want to encourage people who think they can't, because I am proof they can.