HPCC64 is a human–AI research and engineering lab focused on complex technological, business, and social systems. We study how to understand systems as they actually operate, make robust decisions under uncertainty, preserve the evidence behind those decisions, and turn them into controlled action.
The lab combines human judgement with machine-scale research, analysis, synthesis, and execution support. The objective is not autonomous AI; it is a resilient decision system in which authority, evidence, uncertainty, and accountability remain explicit.
HPCC64 uses a stable principal–agent model. The human and the AI perform different functions; the boundary between them is intentional.
Additional models, agents, tools, and execution environments are temporary bounded capabilities. They do not form a permanent virtual staff and do not independently redefine the objective or authority model.
Our interests meet at one question: how can a complex system create value and remain resilient when information is incomplete, actors have different incentives, and failures are normal rather than exceptional?
Mission-critical infrastructure, business continuity, recoverability, data lifecycle, hybrid architectures, dependencies, and failure modes. The purpose is to design systems for degraded conditions, not only for the happy path.
Compliance, technology due diligence, architecture, lifecycle economics, and transformation decisions. The focus is the gap between what a system is required to do, what evidence shows it actually does, and what a feasible target state should be.
Evidence, claims, provenance, competing interpretations, reproducibility, and decision state. Machine-scale analysis is useful only when conclusions can be traced, challenged, and reconstructed.
Persistent cognitive collaboration, temporary agents, permissions, observability, error handling, and recovery. The research target is useful autonomy inside explicit human control, not unconstrained automation.
Actors, incentives, formal rules, actual mechanisms, power, friction, feedback loops, and path dependence. We are particularly interested in systems where declared structure and operational reality diverge.
Open standards, replaceable components, interoperable data, local and cloud execution, and vendor-independent boundaries where practical. Portability is treated as a resilience property rather than an ideology.
The methodologies are not independent brands. They form a decision-to-action chain, with evidence and knowledge management spanning the whole process.
Establishes the Required State: obligations, mandatory controls, essential capabilities, and non-negotiable constraints. It answers what must be true before a solution can be considered acceptable.
Establishes the evidenced Current State: architecture, dependencies, lifecycle condition, operational capability, economics, gaps, uncertainty, and material technical or organizational risks.
Risk-Guided Technology Advisory (RGTA) is the technology-domain predecessor of the broader, developing Risk-Guided Value Engineering (RGVE) method. RGVE compares feasible changes and target states through MTQR — Money, Time, Quality, and Risk / Uncertainty, so one attractive dimension cannot silently dominate the decision.
A developing decision methodology for uncertain, multi-actor problems. It decomposes the objective and system, exposes assumptions, maps actors and dependencies, generates materially different alternatives, tests scenarios, compares trade-offs, and synthesizes a robust decision. It is not yet a frozen or universally validated specification.
PASGR-7 is a retained project name; no new acronym expansion is currently adopted. Its historically recovered seven-stage formulation is: Problem → Architecture → Scenarios → Grading → Recombination → Review → Run.The developing execution layer paired with PASGR-7. Once a direction is selected, EXEC converts it into bounded, observable action with explicit conditions, telemetry, success/failure logic, error handling, escalation, and reassessment. Its current model remains subject to further formalization and validation. PASGR-7 answers what is rational to do and why; EXEC answers how to act without losing control of the decision.
The evidence and reproducibility layer across the entire loop. It preserves claims, sources, provenance, context, uncertainty, competing arguments, and reusable decision state so analysis can survive individual sessions, models, tools, and execution environments.
HPCC64 provides the common system frame around these methods. Its six domains are Value & Outcomes; Actors & Capabilities; Processes & Operating Model; Evidence, Data & Knowledge; Cognition, Decisions & Automation; and Technology, Infrastructure & Dependencies. Alternatives are evaluated through the four MTQR criteria.
We use these methods in real social and business problems rather than treating them as abstract frameworks. The specific projects may change; the recurring problem classes are more important.
Testing demand, economics, suppliers, channels, locations, operating constraints, and minimum viable commitments before irreversible capital is deployed.
Designing services around customer value, capacity, quality, economics, dependencies, exception handling, and the realities of daily execution rather than a presentation-layer process diagram.
Working through changing rules, multiple jurisdictions, formal procedures, incomplete information, and the difference between nominal requirements and actual operational pathways.
Analyzing systems where technology, organizations, incentives, human behaviour, and external actors interact — especially where local optimization can create system-level fragility.
If you are working on a complex technology, business, or socio-technical problem where evidence is incomplete, constraints conflict, and important decisions are costly or difficult to reverse, we are interested in serious research exchanges and practical collaboration.