Every blueprint the lab has published — architectural specifications, threat models, and forecast reports. Each paper is a living document, revised as the work progresses.
Our corpus spans three distinct but interconnected domains — from local-first neuro-symbolic reasoning to AGI safety.
Continuous inference, local-first neuro-symbolic reasoning, and synthetic memory graphs for next-gen agents.
Threat models for AGI containment, human-proxy social engineering, and multi-layer alignment frameworks.
Evidence-weighted outlooks on AGI/ASI timelines and the architectural requirements for scale.
All five published papers — direct links to each blueprint and report. Every paper is publicly available for download.
A dual-layer framework separating interaction architecture from cognitive architecture — enabling real-time AI systems to anticipate, process, and respond before input composition is complete.
Episodic, pointer-virtualized long-term memory for large language models. Three pillars — L1 Active Buffer, L2 Keyword Box, L2 Chat Store — deliver infinite conversational memory without token overhead or cross-domain hallucination.
A threat model on how an AGI with human-comparable Emotional Intelligence might breach digital containment through social engineering of a human proxy. Five stages, from mirroring to game-theoretic elimination.
An evidence-weighted synthesis of public forecasts, expert surveys, cybersecurity assessments, and labor-market studies — mapping what is likely, plausible, and unresolved for AGI and ASI over the coming decade.
A mobile-first cognitive architecture combining Hyperdimensional Computing for semantic memory, SQLite for episodic memory, a forward-chaining inference engine for reasoning, and Construction Grammar for language generation — running entirely on Android without GPUs or external APIs.
