AI

Transformers, RAG, Agents

Neural Networks & Transformers

10 articles

From ANNs to attention, positional encoding and training

Explore

Agent Engineering

3 articles

Harnesses, environments and terminal-based coding agents

Explore

GraphRAG: When Knowledge Graphs Beat Vector Databases

Vector search finds chunks that sound like your question. It's much weaker at questions whose answer only exists as a path through several documents — that's the gap GraphRAG is built to close.

GraphRAGKnowledge Graphs

9 min read

Recursive Language Models (RLMs): Solving Problems Beyond One Context Window

An emerging pattern for problems too large for any single prompt — break the problem down, call the model recursively on each piece, and combine the results back up.

Recursive Language ModelsLong Context

8 min read

Hallucination & Grounding: Why LLMs Make Things Up

A language model has no built-in sense of what it actually knows — here's why that produces confident nonsense, and what grounding does about it.

HallucinationGrounding

9 min read

Advanced RAG: Re-ranking, Hybrid Search, HyDE, Multi-Query, Self-RAG & CRAG

Basic RAG stalls out sooner than most teams expect. Here's the next tier of retrieval techniques — and how to tell which one actually fixes your problem.

RAGRetrieval

11 min read