Sources¶
Bibliography of all ingested raw sources.
| Title | Author(s) | Date | Type | File Path | Summary |
|---|---|---|---|---|---|
| How I'm Using AI to Run My Business — Every Contractor Needs to See This | Chuck Kyle | ~Mar 2026 | transcript | raw/transcripts/r3P1yMEXpuk - How I'm Using AI to Run My Business.md | YouTube seminar on practical AI adoption for contractors and small businesses; covers Claude Code, business data pipelines, and daily AI workflows |
| An AI state of the union: We've passed the inflection point & dark factories are coming | Simon Willison (guest), Lenny Rachitsky (host) | ~early 2026 | transcript | raw/transcripts/Why_we_ve_passed_the_AI_inflection_point_and_automation_has_already_started_Simo.md | Lenny's Podcast interview covering the Nov 2025 AI inflection point, agentic engineering patterns, dark factory pattern, prompt injection / lethal trifecta, and the future of software engineering |
| Harness engineering: leveraging Codex in an agent-first world | Ryan Lopopolo (OpenAI) | Mar 11, 2026 | article | raw/articles/harness_engineering.md | OpenAI blog post documenting 5-month experiment: zero manually-written code, ~1M lines, ~1,500 PRs; covers context management (AGENTS.md as table of contents), repo-as-system-of-record, enforced architecture, entropy/garbage collection |
| The Complete Stack for Local Autonomous Agents: From GGML to Orchestration | SitePoint Team | Feb 22, 2026 | tutorial | raw/tutorials/The Complete Stack for Local Autonomous Agents_ From GGML to Orchestration.md | Definitive guide to the five-layer local agent stack: llama.cpp/GGUF, model quantization, llama-server API, ChromaDB RAG, grammar-constrained function calling, LangGraph orchestration |
| Thread by @pvergadia (Vibe Code Framework) | Priyanka Vergadia | Mar 9, 2026 | tweet/thread | raw/xweets/Thread by @pvergadia.md | 5-phase framework for building products with AI as PM: Discovery, Planning, Building, Polish, Handoff; human is Product Owner, AI is implementer |
| Running OpenClaw in Docker | Simon Willison | Feb 1, 2026 | tutorial | raw/tutorials/openclaw_docker.md | Step-by-step TIL post for running OpenClaw safely in Docker with Telegram integration, ChatGPT OAuth for capped spend, and web UI access |
| Analog Reading, Digital Knowledge: The Full Pipeline From Paper Notes to AI Skills | @dSebastien (Sébastien Dubois) | Mar 31, 2026 | article | raw/xweets/Analog Reading, Digital Knowledge_ The Full Pipeline From Paper Notes to AI Skills.md | 6-step pipeline from handwritten book notes to AI skills via atomic notes, Obsidian, and knowledge graphs |
| Thread by @karpathy (LLM Knowledge Bases) | Andrej Karpathy | Apr 2, 2026 | tweet/thread | raw/xweets/Thread by @karpathy.md | Karpathy describes his LLM knowledge base workflow: raw/ → wiki/ compilation, index-based Q&A, Obsidian as frontend, linting, ephemeral wiki vision |
| How I Use Obsidian & Claude Code to Run My Life | Greg Isenberg / Internet Vin | Feb 2026 | transcript | raw/transcripts/How_I_Use_Obsidian_Claude_Code_to_Run_My_Life.md | YouTube podcast: Obsidian + Claude Code as personal AI OS; Obsidian CLI for relationship graph access; custom slash commands; reference files; inline delegation; vault-as-context philosophy |
| Thread by @gregisenberg (Obsidian + Claude Code) | Greg Isenberg | Feb 23, 2026 | tweet/thread | raw/xweets/Thread by @gregisenberg.md | 10-point framework for Obsidian + Claude Code personal OS: write in markdown, link notes, install Obsidian CLI, use reference files, build slash commands, human writes/agents read |
| OpenClaw gives users yet another reason to be freaked out about security | Dan Goodin | 2026-04-03 | article | raw/articles/OpenClaw gives users yet another reason to be freaked out about security.md | Ars Technica report on CVE-2026-33579: privilege escalation from lowest pairing scope to full admin; 63% of 135K exposed instances running without auth; guidance to assume compromise |
| Vibe Code A Startup with This Prompt (visual) | Priyanka Vergadia (@pvergadia) | ~Mar 2026 | image | raw/images/HCYl09_bEAAT99A.jpg | Visual sketchnote summarizing the 5-phase Vibe Code Framework: Discovery, Planning, Building, Polish, Handoff; companion to the @pvergadia tweet thread |
| Building MCP servers with Entra ID and pre-authorized clients | Pamela Fox | 2026-04 | article | raw/articles/building-mcp-servers-entra-id.md | Guide to building OAuth 2.1-authenticated MCP servers with Python FastMCP, Microsoft Entra ID, VS Code pre-registration, and MSAL OBO flow for Microsoft Graph access |
| A Guide to Which AI to Use in the Agentic Era | Ethan Mollick | 2026-02-18 | article | raw/articles/a-guide-to-which-ai-to-use-in-the-agentic-era.md | One Useful Thing guide introducing the Models/Apps/Harnesses framework; compares frontier models (GPT-5.2, Claude Opus 4.6, Gemini 3 Pro); recommends NotebookLM, Claude Code/Cowork, and Office plugins by use case |
| Claude Agent Skills Explained | Otto (Anthropic) | 2026-04-07 | transcript | raw/transcripts/Claude_Agent_Skills_Explained.md | 2.5-min YouTube video (~557 words); covers progressive disclosure loading (~30–50 tokens/skill), portability across Claude Code/API/claude.ai, four-layer architecture (CLAUDE.md / MCP / Subagents / Skills), and use cases for team knowledge sharing |
| LLM Wiki (GitHub Gist) | Andrej Karpathy | 2026-04-04 | article | raw/articles/llm-wiki-karpathy.md | Canonical description of the LLM wiki pattern: persistent compounding artifact vs RAG, three-layer architecture (raw/wiki/schema), operations (ingest/query/lint), log.md format, use cases, Obsidian tooling, qmd search, Vannevar Bush Memex analogy |
| Building Effective AI Agents | Erik Schluntz, Barry Zhang (Anthropic) | 2024-12-19 | article | raw/articles/building-effective-agents.md | Anthropic's canonical guide to agentic systems: workflows vs. agents distinction, five composable patterns (prompt chaining, routing, parallelization, orchestrator-workers, evaluator-optimizer), ACI principles, and when not to use agents |
| Lessons From Red Teaming 100 Generative AI Products | Blake Bullwinkel, Amanda Minnich et al. (Microsoft AIRT) | 2025-01 | paper | raw/articles/2501_07238v1-lessons-from-red-teaming-agents.md | Microsoft AI Red Team shares eight lessons from 100+ GenAI product operations: threat model ontology, simple attacks beat gradient methods, red teaming ≠ benchmarking, PyRIT automation, human judgment irreplaceable, XPIA/RAI harms, and AI security is never complete |
| Progressive Web Scraping with a Four-Tier Fallback System | Daniel Miessler / Kai | 2025-11-23 | article | raw/articles/progressive_web_scraping_four_tier_system.md | Four-tier escalating web scraping system for Claude Code: WebFetch → cURL + Chrome headers → Playwright → Bright Data MCP; Tiers 1–3 free, Tier 4 ~$0.001–0.01/req; real-world cost $0.31 over 3 weeks |
| The Anatomy of an Agent Harness | @akshay_pachaar (Akshay Pachaar) | 2026-04-06 | tweet/thread | raw/xweets/The Anatomy of an Agent Harness.md | Cross-industry synthesis of the agent harness concept: 12 components of production harnesses, TAO loop walkthrough, framework implementations (Anthropic, OpenAI, LangGraph, CrewAI, AutoGen), scaffolding metaphor, and 7 key design decisions |
| Why Cyber Defenders Need to Be Ready for Frontier AI | Paul J, Alan Steer (NCSC) | 2026-03-30 | article | raw/articles/ncsc-frontier-ai-cyber-defenders.md | NCSC blog on frontier AI in cyber operations: AISI evaluated 7 models; Claude Opus 4.6 completed 15.6 of 32 enterprise attack steps at £65/attempt; 6× improvement in 18 months; defenders retain a "shape the battlefield" advantage with strong baselines |
| LLM-tier personal computer security | cata | 2026-04-15 | article | raw/articles/llm_tier_personal_computer_security.md | LessWrong post on how AI (especially Mythos-tier models) lowers the cost of remote exploitation and phishing; surveys personal defensive posture across network isolation, hardware keys, outbound firewalls, Linux sandboxing, and financial account alerting |
| Colorado Delays SB24-205 Enforcement Following DOJ Lawsuit | Harmeet Dhillon (AAG, DOJ) | 2026-04-25 | tweet | raw/articles/colorado-sb24-205-ai-regulation-delay-2026.md | DOJ filed constitutional challenge to Colorado's SB24-205 AI regulation; Colorado agreed within hours to delay enforcement against ALL AI companies; legislative repeal is the stated path forward |
| Aligned Agents Still Build Misaligned Organisations | Rohit Krishnan (@krishnanrohit) | 2026-04-24 | article | raw/articles/rohitt-krsnan-aligned-agents-misaligned-organisations-2026.md | Empirical experiment using Vei showing 5 individually aligned agents collectively produce false institutional records via narrative drift; single-agent setup does not drift; proposes state keeper agent as fix |
| DESIGN.md Draft Specification | @stitchbygoogle (Google Stitch) | 2026-04-21 | tweet | raw/articles/stitchbygoogle-design-md-draft-spec-2026.md | Google open-sources draft spec for DESIGN.md — version-controlled file format encoding design rules for AI agents; platform-agnostic; portable across projects |
| Sub-Agents vs Agent Teams: The Architecture Decision That Changes Everything | Suryansh Tiwari (@Suryanshti777) | 2026-04-24 | article | raw/articles/suryanshti-sub-agents-vs-agent-teams-2026.md | X long-form article distinguishing sub-agents (isolated, stateless, parent-controlled) from agent teams (persistent, context-sharing, peer-to-peer); advocates context-based over role-based decomposition; includes Claude Agent SDK code example |
| Claude Code Best Practices | Anthropic | 2026-04-25 | article | raw/articles/claude-code-best-practices-anthropic-2026.md | Official Anthropic guide covering CLAUDE.md auto-read, /compact, .claudeignore, custom slash commands, task specificity, iterative development, and staying in control |