frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

GitHub appears to be hiding repo stars on mobile for signed-out users

1•ramoz•2m ago•0 comments

Garrett Langley of Flock Safety on building technology to solve crime

https://cheekypint.substack.com/p/garrett-langley-of-flock-safety-on
1•hhs•2m ago•0 comments

Kafka 101

https://highscalability.com/untitled-2/
1•medbar•3m ago•0 comments

Show HN: MCP server that finds dev tool credits in your workflow

1•janaksunil•4m ago•0 comments

Helix: A post-modern text editor

https://helix-editor.com/
3•doener•5m ago•0 comments

Turns out making games is the easy part

1•clamlotus•6m ago•0 comments

Show HN: A governance pattern for self-evolving AI skills

https://github.com/191341025/Self-Evolving-Skill
1•tiansenxu•8m ago•0 comments

Follow-Up: Build Awesome's Kickstarter Is Cancelled

https://brennan.day/build-awesomes-kickstarter-is-cancelled/
2•brennanbrown•9m ago•0 comments

London tech ecosystem map (235 companies)

https://www.londonmaxxxing.com/
2•birdmania•9m ago•1 comments

Polymarket Removes Betting Market on Nuclear Detonation

https://www.wsj.com/livecoverage/stock-market-today-dow-sp-500-nasdaq-03-04-2026/card/polymarket-...
1•bookofjoe•9m ago•0 comments

Agent Office – Slack for (OpenClaw Like) AI Agents

https://github.com/baturyilmaz/agent-office
1•arbayi•10m ago•1 comments

WebSocket+Huffman vs. SSE+JSON for streaming LLM tokens

https://github.com/vidur2/token_entropy_encoder
1•vidur2•12m ago•1 comments

Learning makes brain cells work together, not apart: study

https://www.rochester.edu/newscenter/learning-makes-brain-cells-work-together-not-apart-694722/
1•hhs•12m ago•0 comments

Show HN: WTF-CLI – An AI-powered terminal error solver written in Rust

https://github.com/JitseLambrichts/WTF-CLI
1•JitseLambrichts•12m ago•0 comments

GoldRush Agent Skills for blockchain data and pricing

https://goldrush.dev/agents/
1•Ferns765•13m ago•0 comments

Show HN: Kaeso – infrastructure for connecting AI agents to real services

https://kaeso.ai
1•devinoldenburg•14m ago•0 comments

Telemetry helps. you still get to turn it off

https://ritter.vg/blog-telemetry.html
1•birdculture•14m ago•0 comments

Ask HN: Anyone else feel this community has changed recently?

2•kypro•14m ago•0 comments

Food fortification is cost-effective in fighting hunger across 63 countries

https://www.cochrane.org/about-us/news/food-fortification-highly-cost-effective-fighting-hidden-h...
1•hhs•15m ago•0 comments

Scalify.ai – Order your own website in under 10 minutes

https://www.scalify.ai
1•josh-ternyak•16m ago•0 comments

Supa Claw

https://github.com/vincenzodomina/supaclaw
1•rmason•17m ago•0 comments

Show HN: PureMark – JSON, Base64, Diff tools that never send data to a server

https://json.puremark.app
1•tommy_worklab•21m ago•0 comments

Show HN: An OTLP observability plugin for OpenClaw AI agents in Grafana

https://github.com/awsome-o/grafana-lens
1•AwesomeO3000•21m ago•0 comments

A simplified PostgreSQL-backed ordered message queue with webhook delivery

https://github.com/alaminopu/pypgmq
1•alaminopu•23m ago•0 comments

The Mirror and the Gift

https://www.robpanico.com/articles/display/?entry_short=the-mirror-and-the-gift
1•retrocog•25m ago•1 comments

CERN upgrade to LHCb experiment threatened by UK funding cuts

https://physicsworld.com/a/cern-upgrade-to-lhcb-experiment-threatened-by-uk-funding-cuts/
1•elashri•25m ago•0 comments

Poet-X Enables Billion-Parameter LLM Training on Single H100 GPU

https://www.simplenews.ai/news/poet-x-enables-billion-parameter-llm-training-on-single-h100-gpu-ktw3
1•goldkey•27m ago•0 comments

Show HN: A simple habit tracker focused on clarity and consistency

https://play.google.com/store/apps/details?id=com.shadowstudios.habitonlinetracker&hl=en_US
1•Belsonsan•28m ago•0 comments

Show HN: Kaeso: an OAuth hub for AI agents

1•devinoldenburg•28m ago•0 comments

Show HN: WebBridge turns any website into MCP tools by recording browser traffic

https://github.com/jalabulajunx/WebBridge
2•nonstopnonsense•29m ago•1 comments
Open in hackernews

Can BM25 Be a Probability?

https://softwaredoug.com/blog/2026/03/06/probabilistic-bm25-utopia.html
4•Tomte•6h ago

Comments

PaulHoule•6h ago
When you convert BM25 or other classical IR scores to a probability with this toolbox:

https://scikit-learn.org/stable/modules/calibration.html

you find a lot of things that are unsatisfying such as you never got a relevance score better than p=0.7 or so and that's very very rare. There are many specific problems in IR for which that kind of probability would be really helpful such as combining results that came from different sources or returning a stream of new documents from a collection but it was an early decision in TREC to not reward ranking functions that were good probability estimators or even that are good at the top-1 or top-3 positions but rather reward them for still being enriched in relevant results when you go deep (like 1000 results deep) into the results.

softwaredoug•6h ago
Interesting! I did not know this about TREC's decision or the scipy calibration module.