Google Photos Slides: The Unreasonable Effectiveness of BM25 for Agentic Search
Source
- Author: boxofchocolates
- Source type: Google Photos album import supplied to the wiki builder.
- Imported source layer:
raw/sources/google-photos-slides/aie-slides-9gWZzS1EpXM1C5eK6/ - OCR engine: RapidOCR over downloaded Google Photos images.
Relationship To World's Fair 2026
These photos are matched to the following scheduled session(s):
- 2026 06 29 jo kristian bergum the unreasonable effectiveness of bm25 for agentic search - The unreasonable effectiveness of BM25 for agentic search
Match Confidence
- Confidence: high
- OCR names HORNET, BM25, agentic retrieval, Jo Kristian Bergum, and Hornet.dev.
- The concepts match the official BM25 for agentic search session.
- Timestamp note: Photo timestamps should be treated as capture provenance; official schedule fields remain canonical.
Photo Slides
Photo 008
- Captured at UTC: 2026-06-30T18:26:16.917000+00:00
- Captured local clock: 2026-06-30T11:26:16.917000+00:00
- Raw OCR:
raw/sources/google-photos-slides/aie-slides-9gWZzS1EpXM1C5eK6/ocr/photo-008.txt
OCR text:
HORNET
We bet on BM25.100M web documents.top-
100
engineA
brigh
Photo 009
- Captured at UTC: 2026-06-30T18:26:44.850000+00:00
- Captured local clock: 2026-06-30T11:26:44.850000+00:00
- Raw OCR:
raw/sources/google-photos-slides/aie-slides-9gWZzS1EpXM1C5eK6/ocr/photo-009.txt
OCR text:
HORNET
World'sFair
Four claims to 16/takeaways
remember.
01 The new user Which BM25? 2
GPT-5 searches with operators,years,phrases,retries parser/chunks/fields/k1/b/top-k
E0 04
Why it is effective BM25 +grep is all you need
exact, cheap,inspectable,steerable in the Loop retrieve,inspect,narrow,verify.repeat
Photo 010
- Captured at UTC: 2026-06-30T18:27:50.402000+00:00
- Captured local clock: 2026-06-30T11:27:50.402000+00:00
- Raw OCR:
raw/sources/google-photos-slides/aie-slides-9gWZzS1EpXM1C5eK6/ocr/photo-010.txt
OCR text:
HORNET
World'sFair
18/contact
Follow-up Come talk AgenticRetrieval HORNET
retrieval with me. Edit profile
x.com/jobergum JoKristian Bergum @jobergum
[email protected] Hornet.dev CEO Homet.dev-The retrieval engine built foragents. Trondheim,Norge @hormet.dev Joined July 2009>
1,509Following 20.1KFollowers