How Valgravia was designed — deliberately, then corrected

1. What we believed

That a company’s identity should be designed after its worldview, not before — and that if we reasoned carefully enough from first principles (what is a resource, what is progress, what is judgment), the institution’s constitution, products, and public identity would emerge as consequences rather than inventions. 1

We also believed that writing everything down first was obviously correct. We were half right.

2. What we did

We designed the institution as documents before writing any product code: a thesis, a constitution of seven articles, a covenant between the institution and its people, twelve shared mental models, and three operating doctrines (research, engineering, product). Then we translated that into a public identity — category, promise (decisions you can defend), voice, color, type, motion — and only then designed this website and its opening film. 2

We did not treat earlier drafts as authoritative. Every document went through adversarial review from at least nine perspectives before being locked.

3. What happened

Two of our strongest instincts failed review.

First, our color discipline was too disciplined: an accent constrained to a “5% of any surface” quota would have rendered the identity nearly monochrome and made the one rule nobody could measure — an unenforceable law, which is how real laws start dying. Second, our first symbol — three layered bars with a red top plate — turned out to be visually adjacent to a hamburger-menu icon at small sizes. We had designed meaning that survived logic but not a 16-pixel browser tab. 3

We also confirmed something important: the discussions changed our minds several times (about efficiency as identity, about product strategy, about the primary symbol). The system that emerged is better than any single draft of it.

4. What changed

The accent law is now role-based (evidence, attention, heritage) instead of quota-based. The primary symbol goes through a context test across real placements before being locked. Laws across the institution now live in a register where amendments are logged with reasons — because we learned that unenforceable rules don’t protect a culture, they erode it.

More broadly: we now treat every institutional document as a hypothesis with an expiry review, not scripture. That is the point of this journal. Expect us to be wrong here, on the record, with dates.