We made truth, possibility, morality, ethics, and liability computable for AI.
Runcible turns open semantic matters into bounded, testable adjudications.
Models search the possibility space. Runcible computes admissibility.
1. We made semantic judgment computable.
We made the conditions of truth, possibility, morality and ethics, legality and authority, and liability explicit enough to compile, test, adjudicate, and terminate computationally. For law and authority, qualification is relative to the relevant jurisdiction, governing authority, evidence, and time. The result is a bounded state: warrant, repair, block, escalate, or undecidable, with a reconstructable account of why.
2. We established the architecture for AI to participate in institutions.
Runcible qualifies what may be warranted. Oversing represents the persistent institutional world in which the qualified result becomes operable: people, roles, permissions, authorities, resources, obligations, contracts, workflows, decisions, actions, consequences, and memory. Qualified results can become inputs to further adjudication across groups, organizations, industries, governments, and federated institutions. That compositional architecture is established; generalized higher-order composition and federation-scale operation remain integration and productization work.
3. We apply falsification before admission to consequential action.
Runcible treats every candidate as a hypothesis. It attempts to falsify the construction, exposes missing dependencies, tests constraints and authority, repairs what can be repaired, and refuses closure when proof obligations remain unsatisfied. The governing objective is to prevent unqualified error from propagating across the institutional action boundary. This is a computational distinction between candidate generation and explicit adjudication, not a claim that all error can be eliminated.
4. We implement semantic computation as a language, compiler, and qualification toolchain.
We are organizationally and technologically like other AI labs. Our distinguishing contribution is making semantic judgment computationally explicit: a language, compiler, and qualification toolchain that tests candidate constructions and determines what survives adjudication.
Open semantic source passes through translation and formalization into RDL, the typed semantic language and intermediate representation, with type and concept catalogs. The semantic compiler makes predicates, dependencies, constraints, authority, consequences, and closure requirements explicit. Executable protocols and test suites drive qualification and proof, producing Decidability Records for retention, reuse, and higher-order adjudication. As SQL turns requests into operations over a governed data model, this architecture turns semantic source into objects whose proof obligations can be tested. Recursive retention and adaptation improve the cycle. Specialized-model work includes dedicated attention nodes and RDL-aligned latent/manifold dimensions intended to compress semantic variation and improve compilation, testing, output, and retention. Current surfaces implement this architecture unevenly; complete formal RDL, generalized composition, and those model modifications must be verified in the exact build before being presented as demonstrated capability.
Foundation models industrialized generation. Runcible industrializes semantic adjudication. Runcible is Semantic Computation Infrastructure: a language, compiler, protocol/test system, and qualification runtime for open semantic matters. Models generate possibilities. Runcible determines what survives qualification. Oversing makes qualified states institutionally operable.
Current maturity
The system is early alpha. Public Runcible is strongest on difficult moral and social semantic questions. The Developer Console exposes the current model configuration, protocols, intermediate work, and output machinery. Runcible embedded in Oversing is currently less capable on the hardest semantic questions and more capable at reading, operating, and modifying institutional state. Enterprise hardening, capability parity, broader domain specialization, and generalized compositional adjudication remain incomplete due to the constraints of our current self-funding.
AI can generate a possibility. An institution still has to know what it may rely upon and how to act.
Runcible has two products. Together they complete the work that generation alone cannot: qualifying what can responsibly be used, then applying it through real people, responsibilities, procedures, state, and history.
Foundation models
Generate possibilities at extraordinary speed.
LLMs can draft an argument, a policy, a plan, a report, a label, a message, or a proposed action. Generation expands what an institution can consider. It does not establish meaning, evidence, responsibility, permission, or consequence.
Runcible works with generated output rather than competing to be another generator.
Runcible AI
Turn information into the strongest form you can responsibly rely upon.
Runcible AI curates context, clarifies meaning, finds the propositions and demands inside an item, exposes missing evidence and assumptions, tests what can be decided, repairs what can be strengthened, and preserves what remains open.
An early integration embeds Runcible in Oversing. Separate enterprise and platform licensing remains part of the commercial plan.
Explore Runcible AI ->Runcible Oversing
Apply qualified knowledge through one visible, accountable institution.
Runcible Oversing connects purpose, programs, people, roles, work, procedures, permissions, resources, financial state, decisions, consequences, and history as one current operating reality.
Use it to expand management capacity without losing visibility, accountability, control, or the ability to learn from what actually happened.
Explore Runcible Oversing ->Follow one of your consequential product claims from generated language to responsible action.
Imagine your organization is preparing to put a consequential claim on a product label. A generator can make the sentence more fluent or persuasive. It cannot, by itself, show what the claim means, whether evidence supports it, who must make the decision, or what the organization will learn from the result.
Make the claim answerable.
- Clarify what the sentence actually asserts
- Identify the evidence, definitions, scope, and reliance burden
- Test what the available record supports
- Return the strongest supportable form
- State what remains open and what would resolve it
Keep the institutional matter whole.
- Product purpose, strategy, and program
- Contributors, work, time, cost, and dependencies
- Evidence, procedures, permissions, and decision state
- Responsible roles, launch, and stakeholder response
- Revision history and retained institutional learning
Better information and shared reality make an institution faster, fairer, more adaptive, and harder to displace.
Manage more without losing the field of view.
Purpose, work, people, resources, performance, financial effects, decisions, and consequence stay visible in relation. Leaders can delegate, intervene, reorganize, and adapt while the outcome can still change.
Contribute through purpose rather than politics.
People work with relevant context, explicit responsibility, measurable goals, legible contribution, reciprocal recognition, positive peer feedback, and contestable evaluation. The organization becomes a fairer and more intelligible place to do consequential work.
Receive commitments an institution can answer for.
Products, policies, contracts, opinions, reports, and decisions can remain connected to evidence, responsible action, performance, and consequence. Information asymmetry has less room to become manipulation, avoidable cost, or unowned liability.
Keep learning as complexity compounds.
Shared reality makes drift visible earlier. Transparent responsibility narrows the space for self-interest and capture. Each qualified decision and observed result strengthens the operating asset rather than disappearing into another disconnected system.
Information quality improves coordination. Coordination improves adaptation. Retained consequence turns responsible action into compounding competitive capacity.
Seven advantages the coupled system is built to deliver.
Real-time operating awareness. Transparency that suppresses capture. Continuous adaptation. An institutional world built for AI. A universal coordination platform. Management actionability that is missing today.
Real-time measurement. Real-time measurement of the institution’s operating state — awareness and reaction while there is still time to act.
Transparency against capture. Real-time transparency and accountability that suppress the four capture failings: myopia, Parkinson expansion, purpose drift, and political capture.
Continuous adaptation. Continuous adaptation in response to market pressures.
Built for AI integration. Built as the institutional world AI needs in order to act — not AI bolted onto disconnected tools.
Platform, not customization. A universal operating platform instead of costly custom coordination systems.
Universal coordination. Organizations all do the same thing slightly differently: coordinate. One platform for the coordination problem.
Management actionability. Management transparency and actionability that institutions do not have today.
Runcible is built to make cooperation more capable, not merely to make AI more fluent.
Runcible Oversing grew from a theory of organizational operation developed through building, acquiring, advising, and managing organizations. Runcible AI grew from a science of cooperation that treats information, reciprocity, practicability, liability, and warrantability as conditions for responsible action.
The products are two halves of one problem. The OS provides the institution’s world and memory. The AI improves what the institution may know and rely upon. Neither half substitutes for responsible people; together they make human and machine intelligence more useful inside a real institution.
The advantage is not only better decisions. It is the ability to grow, adapt, and compete without losing the institutional conditions that make good decisions possible.