Spaces

A parameter-by-parameter checklist for evaluating LLMs in production, from architecture and context to serving, tool use and reliability.

The four ways to customize an LLM, from prompt engineering to RAG, fine-tuning and pre-training, and how to choose between them.

Three interactive resources for seeing inside transformers, from Transformer Explainer and LLM Visualization to Transformer Circuits.

Attention from first principles (Q/K/V, self-attention, multi-head, positional encoding) up to transformer encoders and decoders.

Eleven August 2024 arXiv papers on GraphRAG, coding agents, spiking networks, multi-agent RL and new views of attention.

A reference design for using RAG in trading, from live news and market feeds to generated insights and brokerage order execution.

What retrieval-augmented generation is, explained through its three basic steps of retrieval, augmentation and generation.

How subject–predicate–object triples form the building blocks of RDF knowledge graphs, with Turtle examples and SPARQL querying.

Founder blogs, field guides and short videos I keep coming back to.

My notes from Sam Altman's talk for founders: grit, focus, self-belief and real relationships.

My own mental models, plus the classic ones I use for everyday decisions.

Quotes, phrases and rules for life I have collected over the years.

Small Streamlit apps I built: an AI answer evaluator and a house-vs-market ROI calculator.