Journal Archive — Page 162 of 166

Slack & Zoom on NixOS Workspaces Workflow strategy using virtual desktops (7+1 screens) for focus & muscle memory. Includes easy Slack/Zoom install on NixOS via configuration.nix.
Pythonic Chunk-Yielding Generator to Simulate Long-Running Tasks Implement LLM response streaming in FastHTML/WebSockets via Python generator. Cautious method ensures real-time UI & preserves post-stream JSON parsing.
I Don't Know What to Do Sometimes Simulate a Jupyter Notebook UI flow in FastHTML: Plan uses Card components for cell structure, aiding workflow porting from notebooks to web apps.
API, AI & Human Nuance: The Same Thing Navigating nuance in AI interactions (o1-preview), APIs (FastHTML/JS), & client work. Strategies for managing complexity & effective AI-assisted workflows.
The Port For Real This Time Porting apps to FastHTML/Nix: Refactoring ZeroMQ to WebSockets, JSON to SQLite/MiniDataAPI, integrating Botify API, & setting up dev environment.
LLM Ghost In The Machine Implementing LLM 'Ghost In The Machine': Using AI assist for CRUD updates in FastHTML/HTMX, enhanced logging via decorators, and NixOS helper scripts.
Local AI Awareness Training Enable local LLM to execute CRUD in Python/FastHTML app. Teach JSON commands via system prompt, parse/validate LLM output & execute safely via BaseApp.
Local AI In The App Development Loop Implement local LLM command execution in a Python web framework. Plan covers JSON parsing, intent analysis, table lookups, and persistent LLM memory.
Planning to Win with LLM Plan to modify Python BaseApp framework for LLM awareness: Log user CRUD actions via WebSockets (`chatq`) for real-time context & future RAG integration.
Local LLM Web Framework Integration Plan Plan to integrate local LLM, vector DB (FAISS), & SQL DB into Python web framework via NixOS. Covers LLM function calling, context awareness & RAG.