frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Pentagon concludes Alibaba and BYD have links to Chinese military

https://www.ft.com/content/c80ce7a7-983b-447c-88c2-de5db4cb2e0a
1•Gaishan•39s ago•0 comments

Trump Administration Announces That We Don't Know Where the Sun Goes at Night

https://www.theatlantic.com/newsletters/2026/02/trump-administration-science-climate-change/686008/
1•KnuthIsGod•1m ago•0 comments

Border social media searches and English: Australian conservative migration plan

https://www.smh.com.au/politics/federal/social-media-searches-at-the-border-and-english-language-...
1•KnuthIsGod•6m ago•0 comments

Fake job recruiters hide malware in developer coding challenges

https://www.bleepingcomputer.com/news/security/fake-job-recruiters-hide-malware-in-developer-codi...
2•Gaishan•13m ago•0 comments

Drink Whole Milk, Eat Red Meat, and Use ChatGPT

https://www.theatlantic.com/technology/2026/02/rfk-jr-hhs-ai-chatbots/686007/
2•voxadam•13m ago•2 comments

Show HN: Evolved x86 AVX-512 kernels for NF4 LLM inference

https://github.com/Anuar81/genesis-kernel
1•Anuar81•14m ago•0 comments

DHS has reportedly sent out subpoenas to identify ICE critics online

https://www.engadget.com/big-tech/homeland-security-has-reportedly-sent-out-hundreds-of-subpoenas...
4•ggm•14m ago•1 comments

Show HN: Wapuubot, an open source AI agent in your WordPress admin

https://github.com/MattPearce/wapuubot
1•MattPearce•16m ago•0 comments

'Dark matter, not a black hole, could power Milky Way's heart'

https://ras.ac.uk/news-and-press/research-highlights/dark-matter-not-black-hole-could-power-milky...
2•pavel_lishin•18m ago•0 comments

Ancient Robots

https://code.chuanqisun.com/ancient-robots/
1•stackdiver•20m ago•0 comments

Codex CLI might silently re-route your GPT-5.3 traffic to GPT-5.2 instead

https://twitter.com/banteg/status/2023112622950813824
3•behnamoh•20m ago•0 comments

Show HN: JeffTube

https://jmail.world/jefftube
2•dvrp•20m ago•0 comments

You Should Make Your Own OpenClaw

https://blog.alexboden.ca/occam-claw/
4•alexboden•22m ago•1 comments

Zephyr Real Time Operating System from the Linux Foundation

https://www.zephyrproject.org/learn-about/
4•wewewedxfgdf•26m ago•2 comments

China Floating Turbine Passes Testing and Completes a Grid-Connected Flight

https://cleantechnica.com/2026/02/13/china-floating-turbine-passes-testing-completes-a-grid-conne...
1•tempestn•30m ago•0 comments

Windows 11 update KB5077181 causes boot loops and network issues for some users

https://www.neowin.net/news/windows-11-update-kb5077181-is-causing-critical-boot-loops-for-some-u...
2•exploraz•30m ago•0 comments

GitHub - New repository settings for configuring pull request access

https://github.blog/changelog/2026-02-13-new-repository-settings-for-configuring-pull-request-acc...
2•pabs3•31m ago•1 comments

Show HN: ContextSubstrate – Capture, diff, replay AI agent runs (Git agent work)

https://github.com/scalefirstai/ContextSubstrate
1•scalefirst•31m ago•1 comments

Interzone Magazine Collection Archive

https://archive.org/details/interzone-1982
1•notepad0x90•32m ago•0 comments

Chinese Cars are coming to the US

https://www.bloomberg.com/opinion/articles/2026-02-09/chinese-cars-are-coming-to-the-us-like-it-o...
4•andsoitis•32m ago•0 comments

A Defector Explains the Remote-Work Scam Helping North Korea Pay for Nukes

https://www.wsj.com/world/asia/a-defector-explains-the-remote-work-scam-helping-north-korea-pay-f...
2•sam345•35m ago•1 comments

Top non-ad google result for "polymarket" in Australia is a crypto scam

1•rtrgrd•38m ago•0 comments

Show HN: I gave my AI agent $50 and let it trade on Kalshi

https://twitter.com/JackDavis720/status/2023243630492746062
2•JackDavis720•38m ago•0 comments

The heavy reality of Venezuela's oil

https://www.reuters.com/graphics/USA-VENEZUELA/OIL-RESERVES/mypmqearlpr/
8•gmays•48m ago•1 comments

Minnesota Bathrobe Lady Sam Stroozas of MPR News

https://www.mpr.org/stories/2026/02/13/meet-bathrobe-lady-sam-stroozas-mpr-news
3•jmward01•48m ago•1 comments

We are in the "gentleman scientist" era of AI research

https://www.seangoedecke.com/ai-and-informal-science/
1•gfysfm•49m ago•0 comments

I got tired of babysitting Claude,so I built AI agent that run on my laptop 24/7

3•bagula_ai•50m ago•0 comments

Terry Tao – Machine assistance and the future of research mathematics [video]

https://www.youtube.com/watch?v=zJvuaRVc8Bg
1•breve•50m ago•0 comments

Your pet's microchip may now be useless after chip company goes out of business

https://local12.com/news/local/cincinnati-animal-care-shelter-warns-microchips-may-fail-after-maj...
8•cpeterso•51m ago•1 comments

The great software stock meltdown

https://www.ft.com/content/9e5742e4-adae-4178-9951-7372adeb7ffc
1•petethomas•52m ago•1 comments
Open in hackernews

Show HN: OpenSlimedit – Cut AI coding token usage by 21-45% with zero config

https://github.com/ASidorenkoCode/openslimedit
2•aSidorenkoCode•1h ago

Comments

verdverm•1h ago
I don't think having a tool that rewrites tool descriptions is a good idea.

You should put effort into getting them in a good place and accept the token levels (which is part of that design space).

aSidorenkoCode•1h ago
Every API call sends the full tool schema for all available tools. In a 10-20 step session, you're paying for the same verbose descriptions over and over. Models don't need a paragraph-long explanation of read on the 15th call.

This plugin slims descriptions to one-liners like "Read file content." while cutting 21-45% of token usage. No schema changes, no custom tools. Just trimmed boilerplate as an opt-in plugin.

verdverm•1h ago
> Every API call sends the full tool schema for all available tools.

Only if you are doing it wrong, search >>> summarization

Then the other question, is it deterministic between runs or am I going to get a different summary each session, turn, or toolcall? And depending on that frequency, am I using more token than I save by doing summarization for N tools?

Minimizing token usage is not the goal in of itself, re: the ageless tradeoff of quantity vs quality

For some context, my system prompt is around 5k tokens at the start. I put file contents there read/write/agents.md, which save millions of tokens and seems to work better than making them message parts.

> Just trimmed boilerplate

This is not what I see this tool doing. It's automatically manipulating words in the background that you should put far more care and attention towards. Referring to those words as "boilerplate" you can just throw into a slop machines is revealing

aSidorenkoCode•58m ago
the benchmarks show no degradation in task completion with the shorter descriptions. We're in the age where frontier LLMs don't need instructions on how to read or edit a file.

The descriptions aren't dynamically summarized either. They're static in the plugin, same every call, every session. Zero overhead, fully deterministic.

This has been validated in over 3000 benchmark runs in OpenCode and I ran the entire Exercism Python practice suite (https://github.com/exercism/python/tree/main/exercises/pract...) with and without the plugin with identical results. An initial dataset is shared in the repo.

verdverm•50m ago
Have you made that benchmarking process open so others could reproduce it?

> with identical results

If your results are identical, you should be very sus, something is wrong if this is true. Nothing in agentic is reliable of fully deterministic

aSidorenkoCode•43m ago
Good benchmark results don't mean identical outputs. The task completion rate is the same: both pass the same exercises. The paths the model takes differ, but the end result is the same -> pass the tests

The full benchmarking methodology and tooling will be published alongside the paper.

verdverm•41m ago
you used the word "identical" to describe it, not me

words matter

which is why I still think this is a terrible idea, I don't think it holds up in the general case and would, as a peer reviewer, be inclined to believe there is benchmark filtering that makes for good results.

You should use the same benchmarks everyone else is when you write your paper