Aiden Jae is my wife’s jewelry brand. 100% recycled 9k gold, manufactured by an RJC-certified workshop in Bangkok. I built the entire brand system. Identity, photography direction, Shopify architecture, packaging. Every layer had to say the same thing or the brand would feel incoherent to a stranger landing on the site for the first time.

The photography had to match the typography. The typography had to match the packaging. The packaging had to match the product. The product had to match the story. If the photography said “luxury editorial” and the typography said “handmade craft market,” the brand would feel like two different companies sharing a logo. That’s a registration problem. The layers are out of alignment.

New City is a novel I’ve been writing across dozens of AI sessions. It has a different kind of registration problem. The narrative has layers too. The plot layer, the thematic layer, the character layer, the prose layer. Each one has to align with the others or the reader feels something is off without being able to name it.

[VERIFY: specific structural details of New City’s architecture. Peter has mentioned kishōtenketsu structure and a somatic prose layer but I need him to confirm which details are shareable.]

The transfer between brand and narrative surprised me. I expected them to be completely different disciplines. One is commercial, visual, designed for a market. The other is literary, personal, designed for a reader. But the structural problem at the center of both is the same: multiple independent layers that need to be in register with each other.

In the brand, when the photography direction drifted from the identity direction, you could feel it in the Shopify storefront. The pages didn’t hold together. Each individual element was well-made. The combination produced a room that felt uncertain about what it was.

In the narrative, when the prose style in one chapter drifted from the prose style in another, you could feel it in the reading experience. Each individual chapter worked. The sequence didn’t hold together. The reader’s sense of whose voice this is wavered.

SavePoint was born from this problem. Context kept evaporating between AI sessions when I was working on New City. I’d have a breakthrough about the narrative structure, the session would close, and the reasoning that connected everything would disappear. The next session started from zero. I called it SavePoint because I imagined saving my old games before a boss fight. That survival reflex became a protocol.

The interesting thing is that the tools I built for the brand (registration as a diagnostic concept, layer alignment as a quality check) transferred directly into the narrative work. And the tools I built for the narrative (SavePoint for context preservation, voice governance to keep the prose consistent) transferred back into the brand work and then into everything else.

Transfer goes in both directions. The brand made me better at narrative structure. The narrative made me better at brand systems. Both of them made me better at AI governance. The material changed every time. The operation stayed the same: figure out what the layers are, figure out if they agree with each other, build structure that keeps them in alignment.