Doug Turnbull's local_llm_judge experiment on the WANDS furniture e-commerce search dataset found that forcing an LLM to pick a more relevant product yielded 75.08% precision at 100% recall over 1000 …
Software engineer Doug Turnbull tested the jev system-one model from Typesafe.ai against GPT-5 and GPT-5-mini on query classification using the Wayfair WANDS dataset, treating a predicted category as …
Search professionals should prioritize backend software engineering fundamentals and hands-on experience with LLMs, agents, and embeddings, according to a hiring-trends report by Brian Pedersen cited …
New AI teams often fail by skipping rigorous evaluation, underestimating retrieval complexity, and treating context as mere text chunks rather than rich metadata, according to a consultant who has wor…
Hierarchical Navigable Small Worlds (HNSW), the core data structure behind Elasticsearch, Weaviate, Milvus, QDrant, and Vespa, handles vector updates via delete-then-insertion, but most databases mark…
OpenAI's GPT-5.4-mini can classify e-commerce queries more cheaply by hallucinating fake categories and matching them to real ones via embeddings, according to a developer who shared the technique. Th…
A senior developer argues that brute-force vector search with NumPy can outperform vector databases for datasets up to about 1 million documents, citing benchmarks from an M4 MacBook Pro showing 79.7 …
A series of vector database courses, led by Doug Turnbull and Adam Hevenor, will run the week of August 10, covering how to choose a vector database, improve hybrid search, and build a vector database…
A software engineer argues that brute-force vector search with numpy can outperform vector databases for many use cases, citing benchmarks showing 79.7 queries per second on 1 million 384-dimension em…
Principal Component Analysis (PCA) can reduce the dimensionality of embeddings, shrinking memory usage from 14GB for 9 million 384-dimension vectors to a fraction of that by collapsing redundant dimen…
A software engineer argues that writing code remains valuable in 2026 despite AI agents' ability to generate code, because hands-on coding fosters attention, understanding, and ownership of the system…
Software engineer Doug Turnbull argues that developers should prioritize writing code over specifications when working with AI coding agents, using tests as executable requirements to gradually expand…
A new 'bitter lesson' in search technology reveals that algorithms matter less than incentivizing content creators to optimize for a search engine, as demonstrated by Google's PageRank and now by codi…
Agentic search can be implemented in three distinct ways: retrieval-centric, harness-centric, and model-centric. The retrieval-centric approach relies on building high-quality search to guide agents, …