frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Sonic Pi v5 Released

https://www.patreon.com/samaaron/posts/sonic-pi-v5-166001392
1•samaaron•52s ago•1 comments

Kalshi and Polymarket bets on clinical trials criticized as 'ghastly'

https://text.npr.org/nx-s1-5922530
1•1659447091•4m ago•0 comments

HeroBot

https://herobot.co/
1•ingersol•4m ago•0 comments

Notes Compound, Blogs Compile: When I Update a Note vs. Publish a Post

https://www.ssp.sh/brain/sharing-as-second-brain-note-vs-a-blog-post/
2•zazuke•5m ago•0 comments

Honey, I Shrunk the Agency (Our journey building a software factory)

https://vibegui.com/en/article/honey-i-shrunk-the-agency/
1•gadr90•8m ago•1 comments

AI danger explained for people who don't understand full sentences

https://www.reddit.com/r/OpenAI/s/ASRcUl35lX
1•nemo_captain•8m ago•1 comments

Coldcard Postmortem from MicroPython's Perspective

https://github.com/orgs/micropython/discussions/19588
1•mech4bg•12m ago•0 comments

Tears in Rain Monologue

https://en.wikipedia.org/wiki/Tears_in_rain_monologue
1•tosh•13m ago•0 comments

Ask HN: What observability stack are you using for AI agents in production?

2•varda_62892•14m ago•0 comments

Asking "Was EvoMap's Evolver a reference" for Hermes gets deleted and blocked

https://github.com/NousResearch/hermes-agent/issues/10232
2•sundarurfriend•14m ago•0 comments

DeepMind Says Its AI Can Predict Hurricanes Earlier Than Everyone Else

https://www.wired.com/story/deepmind-ai-model-can-predict-hurricanes-earlier/
2•WithinReason•18m ago•0 comments

The Illusion of Visual Tool-Use: A Causal Audit of Thinking with Images

https://arxiv.org/abs/2608.06270
1•sbulaev•20m ago•0 comments

A Rare 4.5-Hour Interview with Huawei's Chief Scientist for Semiconductors

https://chinatechbite.substack.com/p/a-rare-45-hour-interview-with-huaweis
2•hunglee2•21m ago•0 comments

Jujutsu 0.44.0

https://github.com/jj-vcs/jj/releases/tag/v0.44.0
1•birdculture•21m ago•0 comments

A Mirror of Trumps Posts

https://fakenews.social/
1•calypso•23m ago•0 comments

How to Keep Thinking

https://www.seangoedecke.com/how-to-keep-thinking/
2•vcfgdev•24m ago•0 comments

Building an LLM Agent to Play Dwarf Fortress

https://blog.trine.dev/posts/2026-02-28-df-ai-exp/
1•totetsu•26m ago•0 comments

Regex is (almost) all you need (2025)

https://lookingatcomputer.substack.com/p/regex-is-almost-all-you-need
1•aleda145•27m ago•0 comments

Blaise v0.14.0 is here: macOS, BlaiseGuard, math and more

https://github.com/graemeg/blaise/discussions/211
1•mariuz•27m ago•0 comments

Anthropic AI Mandatory Employee Stock Trading Plans for IPO – This Will Not Go W [video]

https://www.youtube.com/watch?v=zi9LsoKISEc
1•xbmcuser•28m ago•0 comments

LLM intelligence vs. cost per task, Dec 2024–Aug 2026

https://www.reddit.com/r/dataisbeautiful/comments/1vhw02h/oc_llm_intelligence_vs_cost_per_task_de...
2•theanonymousone•28m ago•0 comments

I built a free, instant technical and on-page SEO tool and keyword checker

https://www.sitealyser.com
1•johnnyfwk•29m ago•0 comments

Show HN: Benchmark local LLMs fit for your device specs

https://github.com/SupermodularAI/local_bench
1•andrelago•29m ago•0 comments

Informing HTTP Extension Design with Data

https://mnot.net/blog/2026/linting_the_web
1•jiehong•30m ago•0 comments

Show HN: ToolHoard – Compare power tool prices across major US retailers

https://toolhoard.com/
1•wesman5060•35m ago•0 comments

New Mexico Court Orders Meta to Pay $567M

https://www.techpolicy.press/new-mexico-court-orders-meta-to-establish-567-million-fund-to-abate-...
1•DeepLogin•38m ago•0 comments

I wonder if you could make a useful DOS executable with just emoji

https://oldbytes.space/@gloriouscow/117044736309312697
2•birdculture•40m ago•0 comments

Render of the Ford Fathom: Ford's affordable electric pickup truck at $28,350

https://evwire.com/p/ford-fathom-electric-truck-name-price-28350
1•juurikas•43m ago•0 comments

CrossPoint 1.5 has officially been released

https://www.reddit.com/r/xteinkereader/s/AZbairiJzJ
4•kokada•43m ago•0 comments

Show HN: I built a local multi-agent AI for Android on my phone

https://molfar.nova-hata.com/
1•Gucul_Molfar•44m ago•1 comments
Open in hackernews

Show HN: I built a local multi-agent AI for Android on my phone

https://molfar.nova-hata.com/
1•Gucul_Molfar•44m ago

Comments

Gucul_Molfar•43m ago
Hi HN, I’ve spent the last six months building an Android app during my free time, coding it entirely on my smartphone without a PC (using Python and Kivy). Molfar System is a local multi-agent orchestrator. Instead of a single chatbot, you create a "meeting room" with up to 6 agents (experts with distinct roles, a Moderator, and a Secretary). You provide your own API keys (Google, OpenAI, NVIDIA, OpenRouter), and the agents discuss the task based on their roles, completely locally on your device. The Secretary then compiles the discussion into a structured summary. Why it's in closed testing: I am currently stuck in the infamous Google Play developer purgatory: I need to gather 12 testers for 14 continuous days to get production approval. Since this community understands how to generate API keys and values local execution (zero telemetry, no intermediate servers), I thought this would be the best place to find early adopters. Note: I manually add emails to the Google Play testing track. If you click "Join the testing" on the site, I will add you as soon as possible and you will receive the official opt-in link. I would love to hear your feedback on the architecture, the Kivy performance on mobile, or ideas for new agent roles. Happy to answer any questions!