W-LARQL · DRAFT · ONGOING
LARQL ↗
LARQL explores querying learned systems. This dossier connects the software to questions about model state, representations and what it means to make a model addressable.
CHRIS HAY / THE CONNECTED RECORD
One idea.
Many connections.
A system leads to a question. A notebook leads to a film. Each connection returns to a record.
The graph holds the text and the relationships behind Ask. An editorial connection and an association found in film metadata carry different labels.
252RECORDS
33AUTHORED PASSAGES
235FILM CHAPTERS
1879RELATIONSHIPS
SYSTEMS / NOTEBOOK / QUESTIONS
W-VINDEX3 · DRAFT · ONGOING
VINDEX3 ↗
VINDEX3 is part of Chris Hay’s work on model representation. This dossier records the model-as-database proposition and links to the canonical project.
W-HAUSE · DRAFT · ONGOING
HAUSE ↗
HAUSE is a semantic design system for AI interfaces: communicative acts rendered as statements, instruments and performances. This publication consumes its real components and publishing helpers.
W-MCP · DRAFT · ONGOING
MCP-CLI ↗
MCP-CLI provides a command-line interface for working with Model Context Protocol servers. This dossier connects the interface to tool infrastructure and the wider CHUK work.
N-CONTEXT · DRAFT · OPEN
Context should be abundant. ↗
An open notebook proposition: repeatedly reconstructing earlier state may be one reason context feels scarce. Indexing reusable state is a question to investigate, not an established saving.
N-OPERATOR · DRAFT · OPEN
The operator and the model. ↗
An open thought distinguishing prediction of a model’s trajectory from reproducing its computation. It records a question; no experimental finding is asserted.
Q-FFN · DRAFT · OPEN
What if the FFN is a graph? ↗
Could a feed-forward network be usefully treated as an addressable graph? This question connects model representation to traversal and querying, without claiming an experimental answer.
Q-RESIDUAL · DRAFT · OPEN
Can geometry predict the working set? ↗
Does residual-stream geometry predict which parts of a model will be active or resident together? The existing research note records co-activation as observed and useful locality prediction as untested.
What the sources contain.
241 films have catalogue records. 235 chapters identify topics and starting times; 2 films have indexed automatic transcripts containing 108 passages. A chapter title is not a transcript or proof of a claim.
11 editorial records remain drafts. Their text is searchable with that status intact and can be excluded in Ask. Source references identify cited material; linked websites have not been imported in full.
MACHINE-READABLE GRAPH ↗CONCEPTS / DISCOVERY ASSOCIATIONS
agents
23 records tagged agents. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
The AI Agent Memory That's Impossible to Lie To ↗ SOURCE FILMDeepSeek 3.2 vs GPT 5.2: Agentic AI Face-Off ↗ SOURCE FILMKimi-K2 Thinking Agents TESTED!!! ↗ SOURCE FILMHost your own agents in Zed with Agent Client Protocol ↗ SOURCE FILMCreate a super fast MCP server, deploy and connect to openai agent builder in 15 MINS ↗ SOURCE FILMOpenAI Agent Builder - Beginners Tutorial ↗ SOURCE FILM
ALL 23 ASSOCIATED RECORDS ↗ai hardware
1 records tagged ai hardware. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
ALL 1 ASSOCIATED RECORDS ↗ai interface
7 records tagged ai interface. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
HAUSE ↗ DRAFTMCP-CLI ↗ DRAFTWhat happens after the prompt? ↗ DRAFTMCP Apps: Rich UI for Any AI Client, No Frontend Required ↗ SOURCE FILMLlama.cpp’s New Web UI Is CRAZY Fast! ↗ SOURCE FILMwebassembly system interface (wasi) changes everything.. getting started tutorial using rust & wasm ↗ SOURCE FILM
ALL 7 ASSOCIATED RECORDS ↗coding
12 records tagged coding. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
How Good is Agentic Coding with Qwen Code? ↗ SOURCE FILMClaude Code Hacks: Automate Bash Tool Installs (Game-Changer) ↗ SOURCE FILMClaude Code + Templates: Bootstrap Projects Fast ↗ SOURCE FILMi really want to say goodbye to copilot... ↗ SOURCE FILMGet Started with Palm 2 (bard) API. Coding Tutorial using TypeScript and Vertex AI on Google Cloud ↗ SOURCE FILMFizzBuzz - A history and how it became popular | Coding Tests | Computer Science ↗ SOURCE FILM
ALL 12 ASSOCIATED RECORDS ↗context engineering
1 records tagged context engineering. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
ALL 1 ASSOCIATED RECORDS ↗ffn
2 records tagged ffn. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
ALL 2 ASSOCIATED RECORDS ↗mcp
10 records tagged mcp. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
MCP-CLI ↗ DRAFTWhat happens after the prompt? ↗ DRAFTMCP Apps: Rich UI for Any AI Client, No Frontend Required ↗ SOURCE FILMMCP Client OAuth Is Simpler Than You Think ↗ SOURCE FILMCreate a super fast MCP server, deploy and connect to openai agent builder in 15 MINS ↗ SOURCE FILMQwen3 + MCP ↗ SOURCE FILM
ALL 10 ASSOCIATED RECORDS ↗mixture of experts
51 records tagged mixture of experts. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
Can geometry predict the working set? ↗ DRAFTBetween the facts and the hype. ↗ DRAFTMixture of Experts, Plus One: Injecting Python into GPT-OSS ↗ SOURCE FILMMoE Models Exposed | Position vs Context ↗ SOURCE FILMClaude 4: MoE model? ↗ SOURCE FILMAnthropic reveals hardware specs and Claude updates, OpenAI talks security, and Runway's new model ↗ SOURCE FILM
ALL 51 ASSOCIATED RECORDS ↗model as database
9 records tagged model as database. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
ALL 9 ASSOCIATED RECORDS ↗model memory
8 records tagged model memory. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
The AI Agent Memory That's Impossible to Lie To ↗ SOURCE FILMThe Model Doesn't Unpack Its Memory ↗ SOURCE FILM370,000 tokens loaded in Context in 2.8MB, on a MacBook. ↗ SOURCE FILMWe Don't Need KV Cache Anymore? ↗ SOURCE FILMMoE Models Exposed | Position vs Context ↗ SOURCE FILMHow to Give Your AI Agent Infinite Context with Hierarchical Memory ↗ SOURCE FILM
ALL 8 ASSOCIATED RECORDS ↗model representation
4 records tagged model representation. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
ALL 4 ASSOCIATED RECORDS ↗open models
34 records tagged open models. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
Between the facts and the hype. ↗ DRAFTI Quantized Qwen3.8-27B. “4-Bit” Doesn’t Mean What You Think ↗ SOURCE FILMI Decoupled Attention from Weights - Gemma 4 26B ↗ SOURCE FILMLLMs Are Databases - So Query Them ↗ SOURCE FILMGPT-OSS-20B Has a Secret Layer That Classifies Every Question Before Answering ↗ SOURCE FILMMixture of Experts, Plus One: Injecting Python into GPT-OSS ↗ SOURCE FILM
ALL 34 ASSOCIATED RECORDS ↗quantization
1 records tagged quantization. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
ALL 1 ASSOCIATED RECORDS ↗recurrent depth
1 records tagged recurrent depth. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
ALL 1 ASSOCIATED RECORDS ↗residual stream
2 records tagged residual stream. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
ALL 2 ASSOCIATED RECORDS ↗semantic design
1 records tagged semantic design. These are discovery associations from editorial tags or film metadata, not evidence that the records agree.
ALL 1 ASSOCIATED RECORDS ↗