Weaviate Podcast #141: Drowning in Documents author on why retrieving more and reranking isn't always better
CShorten30 · x · 2026-08-17
Episode 141 of the Weaviate Podcast features Mathew Jacob, now a Ph.D. student at the University of Washington, diving deep into "Drowning in Documents" — the influential work he led at Databricks.
The paper's counterintuitive finding: in reranked retrieval pipelines, retrieving more candidates does not always yield better results, contradicting the common intuition of "retrieve more, rerank, get better." Weaviate calls it a must-read for anyone working in search and information retrieval, and one of their key references while scaling reranked retrieval.
More from Research
- New Dataset: 35k Hugging Face Model Summaries Generated for $0.43 per 1k Rows — vanstriendaniel · 2026-08-18
- The SKILL.md Fallacy: Coding Agents Should Use Ephemeral Subagents — rseroter · 2026-08-18
- FT reveals surge in empirical economics papers as structural models decline — soumitrashukla9 · 2026-08-18
- RLC 2026: Exploring Reinforcement Learning for Fun in EA SPORTS FC 25 — allenainie · 2026-08-18
- Microsoft Research: discarding singleton proteins in PLM training is likely wrong — KevinKaichuang · 2026-08-18
- PKU PhD to present latent space reasoning optimization methods — 青稞AI · 2026-08-18