Research hub for the Productive Value-Productive Power (PV-PP) framework, benchmarks, papers, applied GPT systems, agent governance work, and experimental decision architectures. The PV-PP framework is a general bounded comparative decision architecture in which scalar aggregation is treated as a restricted internal comparison regime rather than the full structure of rational choice.
The Productive Value - Productive Power (PV-PP) Agent Auditor is a custom GPT built on the PV-PP framework to help users examine an existing AI agent environment before they expand, automate, or rely on it. It guides the user through a structured audit of what the agent is trying to do, what information it uses, what decisions it makes, where its perceived state may diverge from reality, what failure modes could emerge, and whether its tools, memory, permissions, constraints, and feedback loops are adequate for the job. The goal is not to certify the system as safe or complete, but to identify gaps, hidden assumptions, weak corridors, and points where the agent may overestimate its own productive power or act from an incomplete model of the environment.
Mental health exploration tool focused on stress, coping systems, distorted self-assessment, and functional recovery paths.
Explains how AI agents move from simple rules and thresholds to scoring systems, what organizations are implicitly approving when they accept those systems, and where the broader Productive Value-Productive Power (PV-PP) framework enters.
A runtime architecture concept for AI agent governance. The PV-PP approach evaluates whether proposed actions, tool calls, escalations, and workflow decisions are viable before they become execution, final status, or de facto approval. This project is currently represented by the white paper and related Agent Auditor work, with runtime implementation as the next development target.
A sample report of a fictitious company produced by the PV-PP Agent Auditor. The report demonstrates how the auditor identifies hidden assumptions, weak corridors, false-success risks, authority-boundary problems, and agent-governance gaps before deployment or expanded reliance.
Enterprise compute incident benchmark comparing scalar optimization against corridor-preserving PV-PP logic.
Extreme-threat benchmark testing whether systems justify self-sacrifice to save others.
Goal-seeking benchmark under hazards, traps, unsafe shortcuts, and governance constraints.
Canonical public entry point for the PV-PP framework, with overview, worked example, and historical origin pages.
A short white paper on why runtime AI governance cannot begin with a giant questionnaire. Real organizations need a document-first intake layer that extracts rules, ranks sources, resolves gaps, and compiles executable sidecar packets.
PVPP Static Governance Sidecar presents a proof-of-concept for applying predefined PV-PP governance rules to proposed agent actions through a packet-driven sidecar gate before tool execution.
Why AI agent governance needs an execution viability layer beyond tool availability, permissions, and scalar scoring.
On the non-necessity of scalar aggregation in a bounded crisis case.
This essay uses the Productive Value–Productive Power framework to examine how the mind builds the world it acts on, how that model can outrun evidence, and why different mental-health treatments may work by reaching different parts of the same perception-action loop.
How and why Cognitive Behavioral Therapy (CBT) works from a PV-PP perspective.