Artificial Intelligence & Machine Learning

Under Technique, Artificial Intelligence & Machine Learning covers methods and practices for building systems that learn from data or mimic intelligent behavior—not specific agent products, model hosts, or IDEs.

In this subcategory: classical and deep learning patterns, training/evaluation discipline, feature and pipeline thinking, and applied ML workflow (experimentation, datasets, metrics). Use the vault hub AI & ML for deeper reading lists and definitions (AI ⊃ ML ⊃ deep learning).

Sibling subcategory: AI Techniques — how you apply models in software (e.g. Agent Skills, gbrain two-layer docs, prompt patterns). Not here: runnable agent tools (cursor-agent, Claude Code → AI Agent under Tool); model platforms (Ollama, cloud APIs → Platform); competitions/community hosts like Kaggle (platform/community, not a technique).

Garden stance: learn fundamentals here; adopt bounded agents and editors (Cursor, cursor-agent) for day-to-day engineering. Prefer explicit pipelines (Dagu) plus targeted LLM steps over undifferentiated “AI everywhere.” See projects/research/morning-briefing-personal-automation-platform for automation boundaries.

Tag an item here when the note is about an ML/AI technique or practice, not a vendor product or agent runtime.

No items published for this subcategory yet.