Ratio et Vis: Two Halves of One Practice
The name is deliberate. Ratio et Vis, in Latin, identifies two distinct disciplines that any sophisticated product practice must contain.
Ratio (from L. reason, calculation, judgment) is the work of thinking before building. The discipline of: should this exist? The output is not a product — it is judgment, supported by data drawn from operation in real conditions. Ratio engages when an idea exists but evidence does not.
The work of Ratio is fast. It must be. The question it answers — is this worth production capital? — has a half-life. Take too long, and the answer matters less. A Ratio engagement that takes six months is not Ratio; it is procrastination dressed as engineering.
Vis (from L. force, strength, vital power) is the work of operating at scale. The discipline of: how does this run continuously, under real load, when it must not break? The output is a system — engineered from first principles, instrumented for stochastic behavior, bounded in cost, mapped for failure modes that will occur. Vis engages when a concept is validated and production capital is committed.
The work of Vis is slow by design. Production systems do not earn their keep by being fast to ship; they earn it by being honest about what production demands. A Vis engagement that ships in two months and breaks in six is not Vis; it is a Ratio engagement that confused itself.
The two modes are sequential, not migrations
When both happen, they happen in order: Ratio first, Vis second. But neither is a migration of the other.
A Ratio engagement is complete on its own. Its value is judgment — the decision to commit, or not commit, production capital. A team can do Ratio with us and never touch Vis, because the answer may be no, do not build this. That is success, not failure.
A Vis engagement is complete on its own. Its value is operation — a system that runs continuously under real conditions. A team can come to Vis with a concept validated elsewhere — internally, by another partner, or by adjacent work. We engineer it; we do not require to have validated it.
Treating Vis as "the next phase of Ratio" is the most common mistake in AI projects. The shortcuts that make Ratio fast (hardcoded data, simplified logic, single-user testing) become liabilities when they migrate to production. The architecture that makes Vis robust (scaled load profile, failure isolation, recovery paths) cannot be retrofitted from a Ratio codebase. They must be built fresh, by people who understand both, from the validated concept up.
Under one practice
What unites the two modes under one practice is structural integrity: the same authorship principle (principals only, no junior bench), the same engineering doctrine (engineering, not magic), the same rule beneath every refusal (the work over the engagement).
The practice is not a methodology to follow. It is a way of seeing the work.
Founders, CTOs, and innovation directors who need ideas validated come to Ratio. Those who need systems engineered for production come to Vis. The two halves are equal. One purpose: sophisticated products that meet their audience and survive at scale.