I spent a year building an AI operating system for creators. Here's what I learned.
For a long time, I believed my biggest challenge was building digital products.
Eventually, I realized I was solving the wrong problem.
The real challenge wasn't creating ideas.
It wasn't writing content.
It wasn't even AI.
The real challenge was keeping everything connected.
Every project I started followed the same pattern.
Ideas lived inside AI conversations.
Research ended up in one document.
Planning happened somewhere else.
Launch notes lived in another tool.
Every new project meant rebuilding the same structure from scratch.
The tools themselves weren't the problem.
The workflow was.
That realization stayed with me for months.
I wasn't looking for another AI assistant.
I wasn't looking for another prompt library.
I wanted something that connected the entire journey—from the first idea to launch.
That idea eventually became MORPHOICES.
Not as a finished platform, but as an evolving operating system for building digital products with more structure and less fragmentation.
Looking back, I think the biggest lesson wasn't about AI.
It was about clarity.
The more connected a workflow becomes, the easier it is to make better decisions.
That's the direction I'm continuing to explore.
I'd love to hear how you organize your own projects today.
What's the biggest source of friction in your workflow?
Thanks for following the journey.
MORPHOICES is being built in public—one idea, one workflow, and one iteration at a time.
If you'd like to explore the philosophy behind the project, you can download the free Blueprint or follow the journey here:
The line that resonated most is "every new project meant rebuilding the same structure from scratch." I build LLM tooling and see this constantly: the expensive part isn't any single AI step, it's that the connective tissue between steps lives in nobody's system of record. One thing I'd gently push on: "operating system" as a framing sets you up to compete with every tool your users already have open. The wedge that worked better for me was picking the single most painful handoff (for us it was prompt-to-deployed-endpoint) and owning just that transition end to end. Curious which specific handoff your early users say hurts most; that answer probably matters more than the connected-everything vision for the next six months.
The shift from “building products” to “keeping the work connected” is an interesting distinction.
That feels like a different problem than simply having too many tools.