Qdrant meetup: better chunking doesn't mean better retrieval, plus four flavors of agent memory
qdrant_engine · x · 2026-10-21
- At MLOps Community Amsterdam, Qdrant's @EASzyszka walked through chunking benchmarks, showing that better chunking does not equal better retrieval, with token-per-correct-answer curves across corpus sizes.
- A second talk covered agent memory, breaking it into working, episodic, semantic and procedural flavors, plus lessons on keeping source consistency in agents. Recap by Luca Berton.
More from coding & agent
- OpenCode hacked, eMudra used private key of 4, PHP hit by supply-chain chaos — burny_tech · 2026-10-21
- Indie hacker ships 8 email product features in one drop with Opus, shares inbox open-rate data — marclou · 2026-10-21
- Solo dev runs 6,000+ unattended news episodes in 19 months — eval loops, not models — jwdNyc · 2026-10-21
- Prototyping physical products with Grok Bot: from sketches to RFQs — elonmusk · 2026-10-21
- A Practical Roadmap for Evaluating Agents on Highly Subjective Tasks — cwolferesearch · 2026-10-21
- Free AI Search course covers Late Interaction, MUVERA scaling and Wormhole vectors — CShorten30 · 2026-10-21