OpenMake: A Bet on Intelligence Staying Plural

OpenMake is a bet on intelligence staying plural.

As frontier intelligence becomes cheaper and more widely available, many capabilities will stop being durable advantages. Anything a model can reproduce, derive, or rent to everyone will become table stakes. In that world, companies cannot compete on access to intelligence alone.

The next generation of enterprises will compete by defining and refining an executable theory of their domain: a living model of how their world works, what good looks like, what fails, and what should be verified.

This matters most in domains that are hard to reproduce from the outside: where data is asymmetric, feedback loops are slow, outcomes are hard to verify, and expertise is earned through experience. These are the domains where competitive advantage can still compound.

OpenMake is a product research lab building tools that turn experience into verifiable environments, so enterprises can gain learning leverage from the work they already do.

These environments span the learning spectrum: from facts and procedures injected into context, to verifiers that guide long-horizon autonomous work, to training signals that turn repeated experience into genuine proficiency.

We are building for a future where enterprises compete on domain intelligence: turning what they know, what they have seen, and what they can verify into systems that learn, improve, and last.