frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

New airliner sets record flying 24 hrs nonstop from Australia to France

https://arstechnica.com/gadgets/2026/07/new-airliner-sets-record-flying-24-hrs-nonstop-from-austr...
2•divbzero•5m ago•0 comments

Show HN: PR review time decreased by N3MO:impact analyzer (15 days free trial)

https://n3mo.vercel.app
2•RajX_dev•6m ago•0 comments

PJM board proposes backstop capacity auction, data center curtailment plans

https://finance.yahoo.com/energy/articles/pjm-board-proposes-backstop-capacity-100932204.html
2•measurablefunc•8m ago•0 comments

China tests largest superconducting fusion magnets ever built

https://theprint.in/india/breakthrough-china-artificial-sun-project-6-5-tesla-magnet/2999321/
3•soniman•11m ago•0 comments

More Tailscale tricks for your jailbroken Kindle

https://tailscale.com/blog/jailbroken-kindle-proxy-tun-modes
5•Error6571•11m ago•0 comments

Show HN: Firemaps Spain – Live wildfire map with wind flow for ES and PT

https://firemapsspain.online/
2•kalamarico•20m ago•0 comments

Manganin: Tools Matter

https://blog.manganin.dev/blog/tools-matter/
2•ai2027•21m ago•0 comments

Israel Is Paying Millions to Train AI Chatbots How to Talk About Gaza

https://www.dropsitenews.com/p/israel-brad-parscale-ai-chatbots-gaza
11•jaykru•23m ago•1 comments

Cross-platform Shadow AI: Find, score, and report unmanaged AI

https://www.npmjs.com/package/@guardion/shadow-ai
2•rflsandroni•26m ago•0 comments

Show HN: Wordknots

https://wordknots.barapa.dev/
2•barapa•31m ago•0 comments

Google DeepMind dismantles Nobel-winning AlphaFold team in strategy shift

https://www.ft.com/content/61b2953d-ee0d-45de-af6e-a9c1cf524b33
3•marcopolis•36m ago•0 comments

Cracking Windows Open: Porting RADV to Win32

https://www.collabora.com/news-and-blog/news-and-events/cracking-windows-open-porting-radv-to-win...
4•zdw•39m ago•0 comments

User Interfaces of the Demo Scene

https://www.datagubbe.se/scenegui/
3•zdw•40m ago•0 comments

Kospi Plunges After Nvidia CEO's Visits Spark 'Huang Curse' Fears

https://www.chosun.com/english/market-money-en/2026/07/29/6FEUZWQT5BG3HMJ3G2RZPHROGM/
6•mapping365•40m ago•0 comments

Encode and Decode any digital text for free

https://www.qrexpress.org/ro/
3•levario•41m ago•1 comments

Self-Hosting has 34× First-Year ROI at 90% capacity

https://www.reddit.com/r/LocalLLM/comments/1v8dv47/what_am_i_missing_selfhosting_kimi_k3_has_34/
3•linxy97•51m ago•2 comments

Protecting My Server and Knowing What It Does with Tailscale and Beszel

2•SyedFazil•52m ago•0 comments

Show HN: Robust Blockchain Infrastructure

https://onfinality.io/en
2•cyrbuzz•53m ago•1 comments

A CLI that shows which sub-agent burned your tokens in one run

https://github.com/rrkher059/token-trace-viewer
2•rrkher059•54m ago•2 comments

Certified in Theory, Broken in Practice: Assumption Gaps in Cryptographic Model

https://arxiv.org/abs/2607.21839
2•sbulaev•1h ago•0 comments

The Future Is Here (2009)

https://retout.co.uk/2009/10/25/the_future_is_here/
2•jruohonen•1h ago•0 comments

MonkeysCode and Capuchin AI: The Agentic IDE That Gives You Total Control

https://monkeyscode.com/blog/introducing-monkeyscode-capuchin-the-agentic-ide-that-gives-you-tota...
1•monkeyscloud•1h ago•0 comments

What Has Happened to Taste?

https://www.esquire.com/style/a71276009/what-has-happened-to-taste/
3•acmnrs•1h ago•1 comments

Kestrel – passive AD audit tool on native C

https://github.com/ssteelfactor-oss/Kestrel
1•steelfactor•1h ago•0 comments

Transformer Transformer: A Unified Model for Motion-Conditioned Robot Co-Design

https://transformer-transformer.github.io/
12•ilreb•1h ago•0 comments

Exit Codes

https://bettercli.org/design/exit-codes/
2•ankitg12•1h ago•1 comments

One transcript, four different GPA views

https://mongol-jimmi.github.io/medcompass-audit/guides/gpa-formulas.html
1•medcompassaudit•1h ago•0 comments

An autonomous fleet to turn your Intent into production systems

https://intentlab.ai/blog/turn-your-intent-into-production-systems
1•jinqueeny•1h ago•0 comments

Show HN: Poople Today – Word Ladders Ending in Poop

https://poopletoday.com/
1•aitooltrek-com•1h ago•0 comments

Show HN: I run 30B 22tok/s, 109tok/s not novel,6GB/16GB RAM overcoming llama.cpp

https://github.com/FedericoTs/quantprobe
1•federicoTXTS•1h ago•0 comments
Open in hackernews

A CLI that shows which sub-agent burned your tokens in one run

https://github.com/rrkher059/token-trace-viewer
2•rrkher059•54m ago

Comments

rrkher059•54m ago
A command-line utility to process OpenInference spans generated from a single agent instance and output: - the number of tokens for each sub-agent separated by category - each step sorted by the monetary price - any context block that was re-sent between steps along with the cost of each re-send

Requires Python 3.10+ and no dependencies. The repository comes with a sample trace for testing purposes.

rrkher059•53m ago
Author here. This is kind of my first project, so I would appreciate any kind of feedback, critical or not. Claude Code wrote the majority of this. But what I really need from you guys is to run it on a live trace and report back what it got wrong. here's the quick summary: kind of.

LangSmith and Langfuse allow ranking on cost per step, but as dashboard widgets aggregating data over many runs over a period of time. Helicone allows doing so if you label your call and have to write your own SQL queries. Phoenix provides cost per span and per project but no middle ground. So if you need to know what made this particular run cost the way it costs, you will still need to read the trace tree manually. None of them detect repeated contexts in subsequent steps, which, in my experiments, is the higher figure. limitations because you will discover them anyway:

Repeat detection works with exact matching and prefix sharing. Mismatched strings in between aren't detected. The token counts in repeat detection are len(text) // 4, not true tokenization. Marked as such in the output. Two model price examples are hard coded. All other input produces n/a instead of guessing. This was tested on the included fixtures and one live trace run, the three node LangGraph case. The outputs of Phoenix, Langfuse, CrewAI, and AutoGen have not been processed by it yet.