retrieval
2 published articles
NLP & ML3 min read
RAG research
BM25 beats agentic RAG when corpora pass 10 million tokens
Agentic search wins on small corpora, but a muset-ai study across 28 nested tiers shows BM25 overtaking it near 10 million tokens. The agent burns 39x more query tokens, and graph RAG stalls in construction.
2026-08-07
AI5 min read
AI Infrastructure
The hidden tax every AI agent pays just got a target painted on it
Nvidia's Nemotron 3 Embed models attack the cost of agentic loops. The 1B variants, not the 8B flagship, are the real story for production deployments that count every token.
2026-07-16