Why this site exists
I've been spending a lot of time working with LLMs — Claude, GPT-4, Gemini — and I keep rebuilding the same mental models from scratch every time a new model drops.
This site is my attempt to fix that. One place for:
- Honest model comparisons on real tasks
- Step-by-step guides for things like RAG, fine-tuning, and prompt engineering
- Tools that run in the browser so you don't need to set up a local environment
- Notes from my own experiments
It'll grow over time. The first guides are already in progress.