Agentic
24 notes
- Agentic Engineering Best-practices hub for building with agents while protecting the engineering bar — quality, specs, verification, architecture, and human responsibility. Vibe coding raises the floor; this raises the ceiling. Doctrine layer: Agentic Engineering, Condensed.
- Agentic Engineering, Condensed The agentic engineering corpus as doctrine, split by half-life: invariants expected to survive far more capable models — taste, specs, verification, understanding — and dated tactics stamped 2026-06 and expected to rot.
- Automatic and Deliberate Work with AI <div class="hub-page-title">
- Automation and the Job Iceberg Sixty years of census data record exactly one job eliminated by automation; the rest were transformed or multiplied, and 60% of today's jobs didn't exist 50 years ago. The augment-or-replace question turns on whether a role can absorb responsibility.
- Claude Fable Operating notes on Claude Fable 5: where it earns its keep, where it fails, and the elicit-first rules for taste-bound work.
- Context Engineering LLMs perform better when the available information is shaped around the task, constraints, examples, and desired output.
- Current Agentic LLM Stack This page documents the models, tools, and workflows used for agentic work on the llm-knowledge-base project.
- Gbrain and Lossless - Persistent Knowledge Across Conversations and Recoverable History Gbrain and Lossless are two complementary memory layers that sit outside (or alongside) the raw context window.
- Hermes Agent Hermes functions as an autonomous operator for your work. It maintains continuity across sessions, writes its own tools, and moves projects forward without constant direction.
- Hybrid Model Workflows, Grok + Hermes As my work on the llm-knowledge-base has grown, I’ve been shifting toward using Hermes (the agent) as the primary actor for working on the knowledge base, while still using Grok for higher-level reaso
- Knowledge Base as Thinking Partner A workflow for using the knowledge base to think, decide, synthesize, and generate next questions instead of only storing notes.
- LLM Knowledge Systems Language models can help a body of knowledge compound when they collect, compile, query, audit, and extend durable files such as markdown.
- Obsidian Dashboard Canonical page: Obsidian Dashboard (Self Management). This Systems path was a near-duplicate and is kept only as a redirect.
- Question Answering Against a Wiki Research answers improve when the existing markdown knowledge base becomes the first context instead of an afterthought.
- Raw to Wiki Compilation Source material becomes more valuable when it is turned into durable, linked, cited wiki pages.
- Software 3.0 Natural-language context and prompts become a programming medium when LLMs can interpret them as instructions, constraints, examples, and executable intent.
- The AI Industrial Revolution When agents do the implementation, the engineer's job stops being "ship output B" and becomes "build the factory that ships outputs B through Z." This Naval Podcast roundtable (Guillermo Rauch of Verc
- Thinking Models Some LLMs improve difficult-task performance by spending more internal computation on problems such as math, coding, and logic.
- Tsumugu-ed: composer maintenance batch complete; backlog recorded Verdict: The deferred Composer-native maintenance pass is done — validation green, zhuyin cross-check clean. Next value is either learner-facing (example-sentence pass) or encoding-facing (phonetic-fa
- Vibe Coding Implementation can move quickly when an AI coding agent handles the low-level construction while the human steers with high-level intent.
- What the Model Names Signal Anthropic's model names are a capability ladder dressed as poetry: the form in each name encodes the scale and kind of composition the model is tuned for, from a three-line haiku to a full opus. Readi
- Wiki Breakdown Pass Missing pages, split candidates, and new connections should be found deliberately before hub pages become overloaded.
- Wiki Health Checks Coverage, consistency, and source discipline improve when the wiki gets periodic LLM-assisted reviews.
- Wiki Status Checks The knowledge base needs lightweight read-mostly audits that answer: what shape is it in right now, and what should improve next?