| // options | |
| String imagePath = "img/desandro-avatar.jpg"; | |
| float res = 10; // resolution, dot density | |
| float zoom = 3.0; // proportion to zoom image | |
| boolean isAdditive = true; // true = RGB, false = CMY | |
| float offsetAngle = 1; // affects dot pattern | |
| /* | |
| 0 = no patter | |
| 0.05 = slight offset | |
| 1 = slight offset |
| // Media Queries in Sass 3.2 | |
| // | |
| // These mixins make media queries a breeze with Sass. | |
| // The media queries from mobile up until desktop all | |
| // trigger at different points along the way | |
| // | |
| // And important point to remember is that and width | |
| // over the portrait width is considered to be part of the | |
| // landscape width. This allows us to capture widths of devices | |
| // that might not fit the dimensions exactly. This means the break |
| /*----------------------------------------------------- | |
| 960 Grid System ~ Core CSS. | |
| Learn more ~ http://960.gs/ | |
| Licensed under GPL and MIT. | |
| -------------------------------------------------------*/ | |
| /* Grid Settings | |
| ---------------------------*/ | |
| // 12-column grid |
| file=/Applications/Dropbox.app/Contents/Resources/check.icns | |
| [ -e "$file" ] && mv -f "$file" "$file.bak" | |
| unset file |
A pattern for building personal knowledge bases using LLMs.
This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.
Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.