Research Notes

Ideas at the intersection of AI infrastructure, cognitive architecture, and the systems that will run autonomous agents.

Featured

The Semantic Mesh: Why Runtimes Aren't Enough

Runtimes make knowledge work executable. But isolated execution doesn't compound. The semantic mesh is the layer that connects structured outputs into accumulated organizational intelligence.

Feb 22, 202614 min read
Featured

The Missing Runtime for Knowledge Work

Agentic AI thrives in software engineering because code already runs inside a repeatable execution environment. Most knowledge work doesn't. The ceiling isn't model intelligence, it's the absence of runtimes that can execute, validate, and provide feedback on structured work.

Feb 21, 202615 min read
Featured

RAG Was Designed for a World Where Intelligence Is Expensive. That World Is Ending.

The economic assumptions baked into RAG architectures are about to become obsolete. And with them, the ceiling they impose on what AI can actually do with your data.

Feb 18, 202613 min read
Featured

Storytelling Through Feature Engineering: Lossy Compression for Language Models

In the LLM era, compression isn't just optimization — it's cognition. How lossy compression with intent turns high-entropy operational data into structured signals that language models can reason over consistently.

Jan 18, 202612 min read
Featured

Query Provenance Store: The Accountability Layer for AI Agent Decisions

When an agent makes a decision, you should be able to see exactly what it saw — and whether reality has changed since. Introducing QPS, a companion spec to TMS.

Jan 17, 20267 min read
Featured

Tabular Manifolds: The Cognitive Interface Layer Between Data and AI Agents

Dashboards were built for eyes. LLMs need something else — a structured, multi-resolution data format optimized for machine cognition. Enter Tabular Manifolds.

Jan 11, 20266 min read

The Missing Layer: Why Agents Need Cognitive APIs

We've been building the wrong abstractions. Agents don't need dashboards or raw data — they need structured reality objects they can reason over efficiently.

Jan 10, 20263 min read
agent-architecturecognitive-systems