I’ve spent most of my career being able to see the whole thing, the story a project is supposed to tell, and watching the process lose it. Not from a lack of talent. I’ve worked with capable designers, engineers, managers. But none of them could see the full picture. Everyone was good at their piece. Nobody held the story.
What comes out the other end of that isn’t bad work. It’s work that doesn’t add up to anything. Decisions that each made sense alone, but never served the same target. The target lived in my head and nowhere else. You can’t blame someone for missing something they were never shown.
So I started building structures that make the target visible. Not for me. I can already see it. For everyone who has to make decisions along the way without losing the thread.
A founder couldn’t make a stranger see the quality through a screen. I built the system that does it. She runs the business. A platform I architected in 2013 still serves organizations you’d recognize, twelve years later. The structures work after I leave the room.
The Savepoint Protocol exists because understanding is temporary. When the context closes, the thread disappears. The Order of the Aetherwright exists because creative identity erodes under scale if it hasn’t been made explicit enough to defend. The household systems exist because my family navigates constraints that make improvisation impossible. The routines hold the decisions because the people can’t. The novel has beat specs and governing constraints because a hundred thousand words will lose the idea if the structure isn’t there first.
Twenty-five years of this: enterprise platforms, brand identities, household systems, fiction, music, AI collaboration. The tools change. The question is always the same: what is fragile here, and what keeps it from breaking?
When I look at a project, mine or someone else’s, the first question is always the same: what is this supposed to be? Not what does it look like, not what technology does it use. What is the thing, at its core, actually supposed to say?
Most projects never answer that clearly enough. Decisions start getting made (about the brand, the photography, the code, the copy) and each one makes sense in isolation. But nobody locked down the essence before the decisions started. So the pieces don’t contradict each other obviously. They just don’t add up.
Once I know what the thing is, the test is simple: does every decision support it? The visual identity. The platform architecture. The photography direction. The packaging. The voice. If any of them contradicts the essence, it’s wrong. Not stylistically, structurally.
What makes this work is that I can move through the domains myself. Design, engineering, and user experience / information architecture. I’ve worked professionally in all three. Most organizations split these across departments. A designer hands off to an engineer who consults a UX strategist. Each one does good work. The joints between them are where the essence gets lost.
I don’t have those joints. When the photography direction contradicts the platform architecture, I can see it, because I built both. When the brand voice says one thing and the information architecture says another, I catch it, because I wrote both. UX and IA are the connective tissue. They’re what turns a visual identity and a codebase into something a person can actually navigate. Without that middle layer, design and engineering talk past each other.
And after years of working at a high level in each, they inform each other. Engineering constraints sharpen design decisions. Visual systems thinking improves how I architect code. Information architecture shapes the UX before wireframes start. In most organizations, these disciplines fight. In my head, they’re symbiotic, and each one makes the others better.
That’s the job. Figure out what the thing is. Then make sure every decision, in every domain, still serves it by the time it ships.
The newest version of that problem: what happens when your collaborator forgets everything between sessions?
It breaks the same way every other collaboration breaks: at the joints. A new contributor doesn’t know what the last one decided. Parallel workstreams contradict each other. The reasoning drifts because nobody maintained fidelity to the original intent. I’ve seen this in design handoffs, in platform teams where the architect left, in brand guidelines that nobody reads. AI just made it visible faster.
The Savepoint Protocol and the Formwork Protocol are structural responses. Savepoint marks where understanding shifts so the thread survives the handoff. Formwork embeds the project’s institutional memory into the project itself so a new contributor reads what’s been decided before they move. Both protocols came out of AI work. Both solve the same problem in any collaboration, because the failure mode is human: continuity destroyed when context doesn’t transfer.