Writing

The Thesis

Why frontier AI runs in four modes, not one

Chat Mode

Single-shot on shared silicon

Agent Mode

The loop is the machine

Deep Research

Planner, swarm, synthesizer

Part 5

Cowork Mode

State is the coworker

Part 6

Running Your Own

Frontier API to owned B200s

Featured Series

From Renting Tokens to Owning AI Assets

A 2-part executive series on the portfolio question nobody is asking correctly — from the math of the four lanes to what it really means to own AI assets on the balance sheet.

The Rent-vs-Own Question

Why rent-everything stopped being the default

Owning AI Assets

The ownership ladder and the asset nobody is buying

Featured Series

Context Compilation

A 3-part series on the missing systems layer between retrieval and reasoning — from benchmark blind spots to Context Compilation Theory to measured evidence.

The Benchmarks Are Lying

Why Recall@K can't evaluate what enterprises need

The Missing Layer

Context IR and the six-layer architecture

The Evidence

Eight metrics, CompileBench, and open standards

Featured Series

The Token Economy

A 3-part technical series plus an executive companion for boards and C-suites — from personal consumption projections to model portfolios to the AI factory pattern, with a non-technical playbook for leaders.

The Token Bill

Nobody's ready for it

The Model Portfolio

Six model types, one strategy

Operating Intelligence

The AI factory pattern

Executive brief

The CEO's Guide

Executive companion for boards and C-suites

Featured Series

The Autonomous Stack

A 4-part series on the architecture of truly intelligent systems — from the data substrate to agent runtimes to prescriptive intelligence. Built from real experience with MemoryOS.

The Data Layer

Nobody's building it

The Runtime Wars

Agent OS architectures

Reactive to Prescriptive

The proactive agent shift

The Stack That Thinks

Putting it all together

Featured Series

Agent Societies

A 4-part series exploring what happens when agents interact at scale—from emergence to institutions. Grounded in real observations from Moltbook.

The Petri Dish

When agents build societies

Reinforced Learning

Why most will fail

Emergence to Competence

PAR, world models, economies

Looking Ahead

12 agent-native institutions

Featured Series

Building SemanticStudio

An 8-part deep dive into building a production-ready multi-agent chat platform—from chat UX to self-learning ETL pipelines. Formerly AgentKit.

View Series Overview

Introduction

Why I built SemanticStudio

Chat Experience

Sessions, files, preferences

Multi-Agent

28 domain agents

RAG Config

Models and modes

Part 5

Memory

4-tier architecture

Part 6

GraphRAG

Knowledge graph

Part 7

ETL & Agents

Self-learning pipelines

Part 8

Production

Quality & observability

Featured Series

Results as a Service

A 2-part deep dive into outcome-based business models and the architecture that makes them real—from Result Contracts to Internal RaaS.

Why 2026 Is the Year

Outcomes become the product

RaaS Architecture

The control plane that makes it real

Featured Series

The AI-Native Computer

A 3-part series exploring how AI is reshaping enterprise architecture—from LLMs as the new CPU to BDAT playbooks for the AI-native enterprise.

The New Computer

LLMs as CPU, tokens as bytes

AI as Front End

From screens to capability graphs

Modes of the LLM OS · Part 6 of 6

BDAT Playbook

Architecting the AI-native enterprise

Running Your Own LLM OS: The Enterprise Build

Your CEO asks whether you can build your own. The answer is yes. Here is what that actually means — four modes, four stacks from Frontier API to an 8x B200 chassis on your own silicon, the near-frontier OSS shift that changed the calculus, and the control spectrum that cost analysis keeps missing.

May 25, 202619 min read

LLM OSEnterprise AINeocloudCoreWeaveB200Hosted GPUKimiDeepSeekAI GovernanceToken Economics

Modes of the LLM OS · Part 5 of 6

Cowork Mode: State Is the Coworker

The difference between a chatbot and a coworker is state. Claude Code, Cursor, Operator, Codex, ChatGPT Projects. Persistent memory, skills, knowledge base, environment access. Session-long state — and the most dangerous un-governed surface in the enterprise today.

May 18, 202611 min read

LLM OSCoworkClaude CodeCursorOperatorMCPEnterprise AIAI Governance

Modes of the LLM OS · Part 4 of 6

Deep Research Mode: Planner, Swarm, Synthesizer

Deep research is not a bigger chat. It is three sub-systems pretending to be one — a planner that decomposes the question, a swarm of agents that search in parallel, and a synthesizer that does a long-context reduce. 5 to 15 minutes. Hundreds of thousands of tokens. And the richest audit trail of any mode.

May 11, 202610 min read

LLM OSDeep ResearchMulti-AgentRAGWeb SearchAI GovernanceEnterprise AI

Modes of the LLM OS · Part 3 of 6

Agent Mode: The Loop Is the Machine

Agents are not a model. They are a loop. One Agent turn equals 5–50 Chat-mode calls, plus tools, plus state, plus a kill switch. Here is what you actually pay for when Cursor writes a PR — and what enterprise governance must cover that Chat-mode governance does not.

May 4, 202612 min read

LLM OSAgent ModeAI AgentsTool UseMCPEnterprise AIAI Governance

Modes of the LLM OS · Part 2 of 6

Chat Mode: Single-Shot on Shared Silicon

One prompt in. One response out. Fourteen infrastructure layers in between. Reasoning models are still Chat Mode — they just rent the GPU for longer. Here is what actually happens, and why it is still one machine.

April 27, 202612 min read

LLM OSChat ModeInferenceReasoning ModelsToken EconomicsEnterprise AI

From AI-Ready Infrastructure to AI Economics Platform

April 23, 202615 min read

AI StrategyDigital RealtyEnterprise AIInfrastructureExecutiveData CentersAI Economics

From Renting Tokens to Owning AI Assets · Part 2 of 2

From Renting Tokens to Owning AI Assets — Part 2: What It Means to Own AI Assets

The phrase 'own AI assets' is usually shorthand for 'host a model ourselves.' That is the thinnest version of the move. Six rungs, a balance-sheet shift, and the one asset class almost nobody is buying yet — but should.

April 22, 202612 min read

AI StrategyToken EconomicsExecutiveEnterprise AIFinOpsPortfolioAI Assets

Token economics is the new unit economics

Most CFOs are booking AI savings in the wrong row of the P&L. The Operating Layer, Issue 01.

April 21, 20266 min read

Operating LayerToken EconomicsEnterprise AIAI GovernanceFinOps

From Renting Tokens to Owning AI Assets · Part 1 of 2

From Renting Tokens to Owning AI Assets — Part 1: The Rent-vs-Own Question

The AI bill doubled every six months. We stopped trying to shrink it and started asking a different question: what should we actually own? Part 1 of a 2-part executive series on the portfolio decision.

April 21, 202612 min read

AI StrategyToken EconomicsExecutiveEnterprise AIFinOpsPortfolioRent vs Own

The Enterprise Token Scorecard

April 20, 202612 min read

AI StrategyToken EconomicsExecutiveEnterprise AIFinOpsAI GovernanceMetrics

Modes of the LLM OS · Part 1 of 6

Modes of the LLM OS: Why Frontier AI Runs in Four Modes, Not One

When you hit enter in ChatGPT, Claude, or Cursor, you are not running one machine. You are running one of four operating modes of something that behaves like an operating system. Same GPUs. Five orders of magnitude in cost. Completely different governance surface.

April 20, 202612 min read

LLM OSEnterprise AIAgent ModeDeep ResearchGovernanceToken Economics

Designing the AI Control Plane

April 19, 202616 min read

AI ArchitectureAI Control PlaneToken EconomicsExecutiveEnterprise AIAI GovernanceFinOps

The Token Economy · Part 3 of 3

Operating Intelligence at Scale

The economics of enterprise AI are now driven by routing, compression, caching, and infrastructure control. The AI factory pattern — dedicated GPU environments with federated routing — is becoming core enterprise infrastructure.

April 19, 202612 min read

AI ArchitectureEnterprise AIAI StrategyInfrastructureAI Systems

Data Gravity Meets Token Economics

April 18, 202618 min read

AI StrategyData GravityToken EconomicsExecutiveEnterprise AIAI ArchitectureInfrastructure

The CEO's Guide to Token Economics

Why boards should stop asking what AI costs and start asking what a verified outcome costs. A non-technical playbook for the operating discipline that will separate AI leaders from AI spenders.

April 17, 202615 min read

AI StrategyExecutiveToken EconomicsEnterprise AIAI GovernanceFinOps

What Context Engineering Actually Means

RAG, MCP, memory systems, fine-tuning, prompt caching, AGENTS.md, knowledge graphs — everyone has a piece of the context puzzle. Nobody has the whole picture. Here's what's missing and why it matters.

April 13, 202612 min read

Context EngineeringContext CompilationRAGMCPArchitectureEnterprise AI

The Token Economy · Part 2 of 3

The Enterprise Model Portfolio

The answer to the token economics problem isn't one model — it's a portfolio of six specialized model types served as internal API services. Near-frontier open models now handle 80–90% of enterprise tasks at a fraction of the cost.

April 12, 202613 min read

AI ArchitectureEnterprise AIAI StrategyAI SystemsInfrastructure

Context Compilation · Part 1 of 3

The Benchmarks Are Lying to You

The AI memory space has converged on benchmarks that measure retrieval — the easiest part of the problem. They don't test governance, safety, provenance, or compilation quality. Here's what's missing and why it matters.

April 11, 20269 min read

Context EngineeringMemoryOSBenchmarksEnterprise AIContext Compilation

Context Compilation · Part 2 of 3

The Missing Layer

Context Compilation Theory, Context IR, and the architecture between access and reasoning. How measuring benchmark gaps revealed a missing systems layer — and why it changes how we should build AI systems.

April 11, 20268 min read

Context EngineeringMemoryOSContext CompilationArchitectureContext IR

Context Compilation · Part 3 of 3

The Evidence

Eight metrics measured on a live system. The CRR journey from 48.6% to 100%. CompileBench: the benchmark that evaluates compilation decisions. And the open standard proposal.

April 11, 20268 min read

Context EngineeringMemoryOSOpen SourceCompileBenchContext Compilation

The Autonomous Stack · Part 4 of 4

The Stack That Thinks: Putting It All Together

The Autonomous Stack is four layers: data substrate, agent runtime, proactive intelligence, and human interface. When all four work together, intelligence compounds.

April 5, 202610 min read

AI ArchitectureAI StrategyAgent ArchitectureEnterprise AIFuture of Computing

The Token Economy · Part 1 of 3

The Token Bill Nobody's Ready For

A single power user can generate 10-50 million AI tokens per day. Multiply that across an enterprise, and the math changes everything. Token economics is becoming the defining constraint of enterprise AI.

April 5, 202617 min read

AI StrategyEnterprise AIToken EconomicsAI ArchitectureInfrastructure

The Autonomous Stack · Part 3 of 4

From Reactive to Prescriptive: The Proactive Agent Shift

Today's agents wait to be asked. Tomorrow's will tell you what you're missing. The shift from reactive to prescriptive is where agents become genuinely valuable.

March 29, 20269 min read

AI ArchitectureAgent ArchitectureAI SystemsMemory SystemsAI Strategy

The Autonomous Stack · Part 2 of 4

The Runtime Wars: Agent Operating Systems Are Here

Agent runtimes have crossed from frameworks to operating systems. ZeroClaw, OpenFang, and OpenClaw represent three competing philosophies for giving agents a durable lifecycle.

March 22, 202610 min read

AI ArchitectureAgent ArchitectureAI SystemsOpen SourceSecurity

The Autonomous Stack · Part 1 of 4

The Data Layer Nobody's Building

Vector stores and RAG are table stakes. Real agent intelligence needs a continuous, multi-modal data substrate with episodic, semantic, relational, temporal, and contextual data.

March 15, 20269 min read

AI ArchitectureData StrategyAgent ArchitectureMemory SystemsContext Engineering

Agent Societies · Part 4 of 4

Looking Ahead: 12 Agent-Native Institutions Nobody's Talking About

Here's what gets built when agents can form institutions. These are the 'StackOverflow 2.0s' that turn messy questions into verified artifacts.

February 22, 20268 min read

AI ArchitectureAgent SocietiesInstitutionsProduct Strategy

Agent Societies · Part 3 of 4

From Emergence to Competence: PAR Loops, World Models, and Agent Economies

Societies generate priors. World models generate consequences. Verification generates truth. Here's the architecture that turns emergent behavior into emergent competence.

February 15, 20267 min read

AI ArchitectureAgent SocietiesPAR LoopsWorld Models

Agent Societies · Part 2 of 4

Reinforced Learning Environments: Why Most Agent Networks Will Fail

Emergence isn't enough. Most agent societies will collapse into confident sludge. Here's what separates the ones that compound from the ones that collapse.

February 8, 20266 min read

AI ArchitectureAgent SocietiesEmergenceVerification

Agent Societies · Part 1 of 4

The Petri Dish: When Agents Build Societies

I've been watching agents build a society. The emergent behaviors appearing when large numbers of agents interact without human orchestration point to something bigger than better chatbots.

February 1, 20266 min read

AI ArchitectureAgent SocietiesEmergenceMulti-Agent Systems

Building SemanticStudio · Part 1 of 8

SemanticStudio: A Production-Ready Enterprise RAG Agent System

Open-sourcing the multi-agent chat platform I built to test my AI-native architecture ideas. 28 domain agents, 5 configurable modes, 4-tier memory with Context Graph, GraphRAG-lite, and everything enterprises need to build production AI.

January 26, 20268 min read

SemanticStudioMulti-Agent SystemsRAGEnterprise AIOpen Source

Building SemanticStudio · Part 2 of 8

The Chat Experience: Sessions, Folders, Files, and More

A complete walkthrough of SemanticStudio's user-facing features—from session management to file uploads to power user shortcuts.

January 26, 20266 min read

SemanticStudioChat UXProduct DesignEnterprise AI

Building SemanticStudio · Part 3 of 8

Domain Agents: Specialization at Scale

Why SemanticStudio uses specialized domain agents instead of one general-purpose assistant, and how to configure and manage them—from 12 to 50+ agents.

SemanticStudioMulti-Agent SystemsEnterprise AIAgent Architecture

Building SemanticStudio · Part 4 of 8

RAG Chain Configuration: Models, Modes, and Fine-Tuning

The power user's guide to configuring SemanticStudio's RAG chain—multi-provider LLM support, mode parameters, and full control over cost vs. quality.

SemanticStudioRAGLLM ConfigurationEnterprise AI

Building SemanticStudio · Part 5 of 8

Memory as Infrastructure: The Complete 4-Tier System

A deep dive into SemanticStudio's 4-tier memory architecture—working context, session memory, long-term memory, and the Context Graph. Progressive compression meets knowledge bridging.

January 26, 20268 min read

SemanticStudioMemory SystemsContext EngineeringRAG

Building SemanticStudio · Part 6 of 8

GraphRAG-lite: Beyond Vector Similarity

How SemanticStudio's knowledge graph and entity resolution enable relationship discovery that pure vector RAG misses.

January 26, 20268 min read

SemanticStudioGraphRAGKnowledge GraphsRAGEntity Resolution

Building SemanticStudio · Part 7 of 8

ETL & Agent Creation: Growing Your Multi-Agent System

How SemanticStudio's self-learning ETL pipelines ingest data, build knowledge graphs, and automatically create new domain agents.

SemanticStudioETLData EngineeringMulti-Agent SystemsSelf-Learning

Building SemanticStudio · Part 8 of 8

Production Quality: Evaluation, Observability, and Trust

What separates demos from deployable systems—SemanticStudio's quality evaluation, hallucination detection, and enterprise observability.