Notes
Thoughts and notes from my digital garden, a collection of ideas growing over time.
Command Line
- awk - Pattern Scanning and Processing Language
- Bash Substitution
Command substitution $(), process substitution <() and >(), and the tee command for piping output.
- Claude Code Bulk Actions
Run Claude Code interactively in tmux popups for batch processing tasks - generating content, updating files, or researching items one by one.
- Gum - Terminal UI Toolkit
Gum provides highly configurable, ready-to-use utilities for building terminal UIs and interactive scripts.
- jq - JSON Command-Line Processor
jq is a lightweight and flexible command-line JSON processor. Think of it as `sed` for JSON data.
- Tmux Popups
Use tmux display-popup for floating windows - quick terminals, music players, AI assistants, and more.
- Well Structured CLI Tool
A guide for building well-structured Bash CLI tools using the dispatcher-namespace-command pattern.
- yq - YAML Command-Line Processor
yq is a lightweight YAML processor. Think of it as jq for YAML data.
- Yq Frontmatter Manipulation
Using yq to read and modify YAML frontmatter in markdown files.
Neovim
- "Open With" Neovim + Ghostty
Automator script used as an "Application" to open files in Neovim via Ghostty on macOS.
- Manipulating Selected Text
- Neovim Basics
- Search and Replace
Searching for patterns and using the substitute command in Vim/Neovim.
- The Quickfix List
A structured list for navigating compiler errors, search results, LSP diagnostics, and other code locations in Vim/Neovim.
- Using Command Line Tools in Neovim
Machine Learning
- Anomaly Detection
Detecting unusual patterns using Gaussian probability models to identify data points that deviate significantly from normal behavior.
- Clustering
- Common Machine Learning Notation Standards
- Gaussian Distribution
- One Hot Encoding
- Overfitting
When a model memorises training data instead of learning generalisable patterns.
- Reducing the number of features
Techniques for determining feature importance, selecting useful features, and combining features through dimensionality reduction.
- Regularisation
Techniques to prevent overfitting by penalising model complexity.
- Supervised vs Unsupervised Learning
- Types of Unsupervised Learning
- Word2Vec
An algorithm that learns word embeddings by predicting words from their context.
Decision Trees
Linear Regression
Logistic Regression
Math
Recommender Systems
- Approximate Nearest Neighbours
Algorithms that quickly find similar vectors in large datasets by trading perfect accuracy for speed.
- Candidate Retrieval
The first stage of a recommender system that quickly narrows millions of items down to a smaller set for ranking.
- Collaborative Filtering
- Hybrid Offline To Online Systems
- Neural Collaborative Filtering
- Online Retrieval And Ranking
- Overview Of Recommender Systems
- Random Walks
A technique for exploring graphs by randomly traversing connections, used to discover item relationships and generate embeddings.
- Recommendation Pipeline
The four-stage funnel that transforms millions of items into a personalised, ordered list for the user.
- User and Item Models
How user and item features are computed offline and combined online to generate recommendations.
Recipes
Wellness
- Quotes
A personal collection of quotes that have stayed with me, gathered over the years.