Joshua O • Feb 11, 2026 • 2 min read • 340 words
A mental model I'm exploring: MVC, not the old timer MVC you know, but the young’un Minimum Viable Constraints
As LLMs commoditize solutions, workflows become disposable. The platform supporting those workflows becomes the product. Our goal becomes density of reliable paths that enables any customer use case (is this the last use case?)
This is an emergence problem. LLMs lower the barrier from idea to execution, and people start combining tools in ways nobody planned for. By definition, you can’t design for emergence, but you can design a system where emergence thrives.
In philosophy, systems theory, science and art, emergence occurs when a complex entity has properties or behaviors that its components do not have on their own, and emerge only when they interact in a wider whole. —Wikipedia
Which presents a new engineering challenge: if your system supports emergence, it also has to handle the outcomes. Not all outcomes will be good. So the question becomes: what's the minimum set of constraints that keeps things working without killing the flexibility that makes the platform valuable?
While MVP strips features to find the core product, MVC strips rules to find the minimum structure that prevents chaos while still letting the system surprise you. We’re moving from a product question to a platform mindset.
Don’t fret, product R&D still matters! You have to solve real problems to learn what the platform needs to support. But increasingly that R&D is what builds the platform, and can be offloaded significantly to your users. When someone tries doing something novel and your platform fails to flex or breaks causality from intent to outcome, that’s free R&D, the backlog and prioritization get automated.
The FDE role is a good example: you embed with customers, solve their specific problem, extract the pattern, and that pattern becomes a new reliable path in the platform. The solution is scaffolding. The platform is what stays.