History — 2026 week 20 (May 11 – May 17, 2026)

Newest first.

2026-05-17 — Consultancy, Tools & Niche Topics (~12 notes)

PeopleCharlie Hull (The Search Juggler, 25+ years in search).

ToolsQuerqy (query rewriting library), Quepid (search relevance evaluation platform).

CompaniesThe Search Juggler.

ConceptsSignal Downboosting.

TopicsUnderstaffed Search Team, Search Consultancy, Events and Conferences, Fun and Philosophy, Books.

ArticlesYour Obsidian Vault Is a Knowledge Graph, Getting Started on Search Relevance for the Understaffed Search Team.


2026-05-16 — Concepts, Companies & Topics Expansion (~49 notes)

ConceptsBERT, Reranking, TurboQuant, RaBitQ, Query Expansion, Collocations, Stopwords, Clicks Residual, MMR, APD, Linear Score Combination, Modality Gap, Contrastive Gap, Conversational Search, Knowledge Distillation, Dimensionality Reduction, Click Models, Query Sampling.

CompaniesQdrant, Netflix, Canva, Carousell, Etsy, Twitter, Spotify, OpenSource Connections, Skyscanner, Grubhub, Cohere.

TopicsA-B Testing for Search, Economics of Search, E-commerce Search, Autocomplete and Autosuggest, Synonyms and Vocabulary Management, Query Understanding in Practice, Managing a Search Team, Search Quality Assurance, Search Result Diversity, Hiring for Search.

PeopleIvan Pleshkov, Jonas Schulz (Qdrant engineers), Neal Lathia (Skyscanner), Stuart Cam (Canva), Hassam Chundrigar (Daraz), Isabella Tromba (Slack), Mihajlo Grbovic (Airbnb), Sergey Feldman (AI2).

ArticlesTurboQuant in Qdrant.


2026-05-15 — Initial Import (~200+ notes)

The founding batch. Core articles, concept definitions, MOCs, and key people were imported from the Awesome Search curated list.

Articles — full Query Understanding series (Query Understanding - Introduction through all sub-topics by Daniel Tunkelang), ranking articles (Multi-Stage Ranking, Compute Mean Reciprocal Rank (MRR) using Pandas), embedding articles (Hypothetical Document Embeddings (HyDE), Fine-tuning Multimodal Embedding Models), and many more.

Concepts — foundational IR concepts: BM25, Reciprocal Rank Fusion, Hybrid Search, Synonyms, Bayesian BM25, Knowledge Graph Search, Query Understanding, Autocomplete, Faceted Search, Spelling Correction, Embeddings, Learning to Rank, NDCG, MRR, RAG, and many others.

MOCs — 6 Maps of Content created: MOC - Search UX and Discovery, MOC - Ranking and Retrieval, MOC - Agentic Search and Embeddings, MOC - Search Quality Assurance and Query Understanding, MOC - Architecture and Search Team.

PeopleDaniel Tunkelang, Doug Turnbull, James Rubinstein, Giovanni Fernandez-Kincade, Wolf Garbe, Eugene Yan, Piotr Mazurek, Karthik Ramasamy, Janani Narayanan, Thomas Veasey, Quynh Nguyen, David Albrecht, Max Irwin, Ted Underwood, and others.