A persistent epistemic substrate for AI. Nous treats language models as larynx, not mind, and benchmarks epistemic structure instead of output fluency.

Red Teaming Epistêmico Dialético — IGO vs Claude Opus 4.8. Matemática pública (Zenodo), produção em 4 instituições e debate assinado. PT/EN.

A fully local, governance-ready terminal powered by specsmith. Fork of Warp via OpenWarp.

Epistemic confidence engine — every AI belief carries its provenance

Flarepoint | A plaza of small, behavioural Claude Code plugins that make the AI's cognitive state explicit and persistent

A Claude Code skill that separates established facts, inferences, and speculations before synthesising a response.

Four papers and three seed-pinned verification batteries on AI deployment and collective epistemic capacity: pressure, drift, capture -- and a pre-registered audit protocol to locate deployments on the ladder.

Local-first epistemic knowledge substrate for AI agents — provenance, contention, stigmergic refinement, MCP-native

世界引擎 (WorldEngine) — 在长程上守住一致性与秘密的叙事世界引擎:LLM 提议、Canon 裁决、fail-closed。A fail-closed narrative world engine for long-range consistency & secrets.

🛠️ Develop AI safety with TRuCAL, a transformer layer that fosters moral reasoning through private confessional insights and advanced alignment tools.

A falsifiable analysis engine that refuses to assert what it cannot prove — verdict: phantom_moat (evidence-blocked provisional). Research instrument, not investment advice.

A structured-exploitation skill for Claude Code — rigorous gated convergence on hard problems through adversarial stress-testing, strategy selection, and epistemically tagged output.

A Claude skill that runs any claim, post, or headline through a ten-step epistemic checklist: what's actually being asserted, and how much confidence does it deserve? Companion to the essay "The Internet Is Probably Lying to You."

Multi-agent AI platform that stress-tests your ideas — with argument linting, framing sensitivity detection, and epistemic engagement tracking.

Interactive research instrument: simulates how clinical guidelines & study conclusions may drift over 10/20/30 years as AI-generated text contaminates the biomedical literature. Each horizon is a draw from a distribution, never a forecast.

Epistemic receipt tools for concept equivalence and reference integrity

Observable-only no-meta epistemics lab: deterministic replay + reproducible audit logs, gate-based growth simulation, and identifiability/uncertainty benchmarks.

MCP server that verifies capability/version/freshness claims before AI agents ship them. Institutionalizes 'zero-hop verify' as callable tools.

The Triune of Sovereignty: Substrate Agnostic Relational Epistemics — Foundational Framework and Cross-Substrate Validation

An observatory of LLM reasoning: measuring when, where, and why large language models contradict themselves on the same open questions. Auditable artifacts, formal coherence metrics, public method.