Company
Why did one company build both an institutional operating system and an epistemic engine?
Runcible turned a science of cooperation into an operating system and an epistemic engine.
One research program. Three organizations. Two products. One institutional capability.
Institutions create extraordinary capability by letting people specialize, coordinate, delegate, and act together. Runcible exists to preserve the conditions that keep that cooperation productive as complexity, intelligence, responsibility, and consequence expand.
The ambition is enabling: let institutions understand more, coordinate more, delegate more responsibly, adapt sooner, and pursue larger purposes without becoming less able to tell what is true or correct what has gone wrong.
Runcible is a long-envisioned infrastructure for civilization: an intelligent partner that produces truth, reciprocity, possibility, and accountability at every scale of human cooperation. The name comes from Neal Stephenson’s The Diamond Age, in which a runcible is the device that carries a child through a continuous education into judgment. What that novel proposed as a thought experiment is what this program is building.
The Operating Origin
The company began with responsibility for operating companies, not an AI market.
How can an owner build, acquire, and manage companies with operational excellence and deep transparency without waiting for delayed, filtered reconstructions of reality?
The question emerged through more than two decades of founding and building consulting companies, serving large enterprises and domestic and international governments, and carrying responsibility for clients, programs, people, delivery, resources, decisions, and financial results.
An owner remains accountable for the whole while causes become local and consequences arrive later. Departments, managers, applications, definitions, incentives, and close cycles each create a partial account.
The problem was not management. Institutions need managers who exercise judgment, coach people, coordinate scarce resources, remove constraints, and accept responsibility. The problem was dependence upon any layer’s privately assembled account of what the institution was doing and why.
The required alternative became one shared institutional model in which purpose, programs, people, dependencies, responsibility, decisions, time, money, action, consequence, correction, and history remain connected.
That operating history is a matter of record rather than recollection. The consulting companies built in that period served startups, mid-market, and Fortune 100 clients, made the INC 500 list three times, were featured in the Forrester Wave, and ranked among the top twenty-five global independent agencies for customer satisfaction.
The software problem appeared only after the institutional problem had been experienced often enough to become a theory.
What Scale Can Break
Opacity allows divergence to become institutional power.
Four pathologies describe what becomes visible. The causal progression explains why it can become self-reinforcing.
The part optimizes its measures while losing the institution’s purpose.
Work and bureaucracy grow without an examinable relation to value.
Groups direct institutional resources toward preferred objectives.
Coalitions control information, procedure, resources, decisions, and incentives.
Runcible does not assume corrupt intent or claim to eliminate human nature. It changes observability, attribution, contestability, and the value of privately controlling the institutional narrative.
An institution remains adaptable only while it can discover and correct divergence from within.
The four failings are the observable surface of capture. The commercial answer is the coupled system’s extraordinary values: real-time measurement and transparency, continuous adaptation, an institutional world built for AI, a universal coordination platform, and management actionability institutions do not have today.
Read the seven pathologies · See the extraordinary-value set
Making Cooperation Testable
A problem this deep required a science of cooperation, not another management doctrine.
The research program worked toward operational definitions and tests for truth, reciprocity, cooperation, ethics, decidability, responsibility, liability, and sustainable institutional order.
What survives examination?
Exact propositions, evidence, tests, repair, and what remains open.
What is supplied and demanded?
Satisfaction, transferred cost, affected parties, and sustainable exchange.
Who must decide?
Role, decision rights, delegation, liability, action, and revocation.
What can actually work?
Constraints, resources, sequence, consequence, and limits.
What can close?
Support, failure, qualification, repair, deferral, or explicit open state.
What keeps the whole coherent?
Purpose, roles, work, judgment, action, consequence, and memory.
Runcible OS
World, purpose, identity, roles, internal law, resources, coordinated action, accounting, consequence, and memory.
Runcible AI
Curation, scientific testing, economic decidability, reciprocity, repair, explicit limits, and warrantability.
The research was not built to explain Runcible after the fact. Runcible was built because the research was intended to operate.
One Program, Three Forms of Work
Science, institutional software, and machine judgment could not honestly be treated as the same job.
The science
- Question
- What makes truthful, reciprocal, decidable cooperation possible?
- Work
- Definitions, reductions, tests, responsibility, liability, and institutional theory.
- Contribution
- Method and constraints supplied to both products.
- Tested by
- Coherence, evidence, adverse cases, and expert challenge.
The institutional world
- Question
- How do the conditions become everyday operation?
- Work
- Programs, structures, roles, workflows, time, goals, communication, finance, and history.
- Contribution
- Runcible OS: world, internal law, accounting bridge, and memory.
- Tested by
- Product behavior, institutional use, and operating result.
The epistemic engine
- Question
- How can information be understood, tested, repaired, bounded, and recorded?
- Work
- Propositions, demands, evidence, scientific method, decidability, open states, and records.
- Contribution
- Runcible AI: embedded capability and separate license.
- Tested by
- Adverse cases, repeatability, integration, and consequence.
Three organizations were the cost of refusing to pretend that science, institutional software, and machine judgment were one problem.
World Before Institutional Intelligence
The operating system revealed what an institutional intelligence would need.
The platform was needed to discover what AI capable of institutional participation would actually have to know, test, remember, and do – and to give judgment a world in which it could acquire accountable consequence.
- ScienceDefine the institutional and epistemic problem.
- Institutional worldRepresent identity, state, work, responsibility, resources, action, and memory.
- Feasible AI methodApply the scientific and economic method through language-native models.
- Integrated intelligenceLet world and mind work upon the same accountable matters.
Historical order describes what happened. World and epistemic mind working together describes what the products are.
Built in Sequence. Coupled in Operation.
Today the program is tangible as two products and one complete system.
Runcible OS
The institutional world and memory. Connect purpose, programs, people, responsibility, resources, work, time, financial state, consequence, and history.
Runcible AI
The epistemic engine. Curate, improve, test, repair, adjudicate, bound, and record consequential information inside a qualified environment.
The Runcible System
AI improves information. Responsible people decide. Authorized action changes state. Consequence returns through accounting, correction, and memory.
In plain language. Runcible produces a governance, constraint, and closure layer that can sit over any AI; a certification service that will test and certify the truth, ethics, legality, and possibility of a claim; and an application platform that individuals, businesses, and governments can build on at any scale. The long-term aim for Runcible AI reaches further than the engine described above: a personal mentor that helps a person become the best they can be from childhood to old age, and in doing so raises the standard of truth, possibility, and cooperation in public life.
Closure Technology
A universal system for testing truth, reciprocity, and demonstrated interests, so an enterprise can be confident its outputs are decidable and defensible.
Training Infrastructure
Methods for converting books, policies, and institutional knowledge into operational training sets that improve models rather than patching them.
Application Platform
AI applications that meet the highest bar of compliance, liability defense, and operational reliability across industries.
Storing information does not make it true. Reaching a supported judgment does not grant the AI responsibility or permission to act.
Disciplines Required by the System
The team spans the forms of judgment Runcible must join.
The people below carry the disciplines the system has to join: institutional theory, medicine, operations and finance, product and engineering, compliance, behavioral science, and law.
Founder track record. B. E. Curt Doolittle founded a dozen companies across technology, law, and research, and built one of the first legal AI systems in the 1980s. He built and sold multiple $100M technology consulting companies serving Fortune 400 clients, including hundreds of millions of dollars of work with Microsoft over thirty years as a Microsoft Solution Provider — at one point the largest privately held Microsoft Solution Provider in the United States. He founded the Natural Law Institute.
The Burden Created by the Origin Story
The longer the program, the greater the obligation to prove it.
A twenty-year narrative is not evidence because it is ambitious or coherent. Each part requires the kind of proof appropriate to it.
Companies, clients, programs, acquisitions, responsibility, and results that may lawfully be disclosed.
Definitions, papers, contrary cases, open questions, and expert review.
Demonstrations, operating records, maturity labels, limits, and corrections.
Repeatability, method boundaries, failures, open states, and integration evidence.
Ownership, licenses, rights, relationships, dates, and corporate authorization.
Customer need, adoption, operating consequence, and bounded economic measures.
Research is what we propose. Products are what we built. Consequences are what will decide whether the program is true.
The Commercial Consequence
The mission advances one institution, one product, and one consequential loop at a time.
Implement Runcible OS to make an institution more visible, coordinated, accountable, and adaptable. License Runcible AI to improve consequential information inside another qualified environment. Use both to build an AI-native institution in which human and machine intelligence learn from the same accountable reality.
Make cooperation durable enough for people and institutions to keep growing.