frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Training language models on TPUs shouldn't be scary

https://dogac.dev/blog/2026/migrating-to-tpu/
1•dogacel•1m ago•0 comments

Show HN: Resilient RAP – A self-healing pipeline for telemetry using embeddings

https://github.com/tarek-clarke/resilient-rap-framework
1•tarekclarke•1m ago•0 comments

India's Bharat Taxi Service as First Cooperative-Owned Digital Platform

https://urbantransportnews.com/news/india-launched-bharat-taxi-service-as-first-cooperative-owned...
1•oumua_don17•2m ago•0 comments

Show HN: Proximity Voice Chat in VS Code

https://nisa.la/vscode-proximity-chat/
1•nkalupahana•4m ago•0 comments

Service Worker, Video Files, and the 206 Partial Response

https://justinribeiro.com/chronicle/2026/02/04/service-worker-video-files-and-the-206-partial-res...
1•speckx•6m ago•0 comments

More time is better than more money (2014)

https://hitotoki.org/moments/dvjugu7d
3•thinkingaboutit•7m ago•0 comments

Examining Turbopuffer ANN v3

https://terencezl.github.io/blog/2026/02/03/case-study-turbopuffer-ann-v3/
1•gk1•7m ago•0 comments

Daydreamers and Sleepwalkers: Crossing the Borderlands of the Unconscious

https://thereader.mitpress.mit.edu/daydreamers-and-sleepwalkers-crossing-the-borderlands-of-the-u...
1•Hooke•7m ago•0 comments

Ubuntu Now Available on SpacemiT K3/K1 Series RISC-V AI Computing Platforms

https://canonical.com/blog/spacemit-announces-availability-of-ubuntu-on-k3-k1-series
1•fork-bomber•7m ago•0 comments

Show HN: Smooth CLI – Token-efficient browser for AI agents

https://docs.smooth.sh/cli/overview
2•antves•7m ago•0 comments

Rising stars of the venture capital industry

https://www.businessinsider.com/venture-capital-industry-rising-stars-list-2026-1
1•cf100clunk•9m ago•0 comments

I replaced QuickBooks with an MCP server running inside Claude

https://tiddwell.com
1•wheeleeo•10m ago•1 comments

I built an AI agent that automatically commented on HN. Here's what I learned

1•Soerensen•11m ago•0 comments

Anthropic Takes Aim at OpenAI's ChatGPT in Super Bowl Ad Debut

https://www.wsj.com/business/media/anthropic-takes-aim-at-openais-chatgpt-in-super-bowl-ad-debut-...
2•Anon84•11m ago•1 comments

Hetzner stopped offering ARM processors

https://old.reddit.com/r/hetzner/comments/1oi1gfk/hetzner_stopped_offering_arm_processors/
2•tosh•12m ago•0 comments

Show HN: How I built Live View for browsers running in Docker

https://docs.blitzbrowser.com/blog/how-i-built-live-view-for-browsers-running-in-docker/
1•sam_march•13m ago•0 comments

Norway only sold 98 diesel cars and 7 gasoline-powered cars in January

https://electrek.co/2026/02/03/even-after-cutting-ev-incentives-norway-only-sold-98-diesel-cars-i...
2•bookofjoe•13m ago•1 comments

In 2026, Postgres Is (Still) Enough

https://postgresisenough.dev
1•cpursley•15m ago•1 comments

OpenAI launches "Frontier," framed as an "HR system for AI agents"

https://www.theverge.com/ai-artificial-intelligence/874258/openai-frontier-ai-agent-platform-mana...
3•unslop•15m ago•0 comments

OpenBSD on SGI (part 1 of 6)

https://miod.online.fr/software/openbsd/stories/sgi1.html
2•zdw•16m ago•0 comments

CryptoHack: A free, fun platform for learning modern cryptography

https://cryptohack.org
1•kblissett•17m ago•0 comments

My review of the Tadpole-RDI Ultra Book lli from 25 years ago this month

https://liam-on-linux.dreamwidth.org/97604.html
2•zdw•17m ago•0 comments

Lego brick with a computer inside (2022)

https://hackaday.com/2022/06/08/computer-shaped-lego-brick-brought-to-life/
1•LorenDB•18m ago•0 comments

Learn prod agentic coding (open source)

https://agenticoding.ai/
2•ofriw•18m ago•0 comments

Show HN: Remove Background for Android

https://play.google.com/store/apps/details?id=com.codegres.removebackground&hl=en_US
1•Codegres•18m ago•0 comments

Go Client for GitHub Actions Runner Scale Set APIs

https://github.com/actions/scaleset
1•crohr•19m ago•0 comments

Show HN: Simple, Fast, Accessible Fine-tuning

https://www.commissioned.tech/
1•rbshamsu•19m ago•1 comments

"The appearance of XXX's name in the files does not imply wrongdoing"

1•chrisjj•19m ago•3 comments

'Ripping' Clips for YouTube Reaction Videos Can Violate the DMCA, Court Rules

https://torrentfreak.com/ripping-clips-for-youtube-reaction-videos-can-violate-the-dmca-court-rules/
3•Signez•20m ago•0 comments

ClawHavoc: Malicious Clawed Skills Found by the Bot They Were Targeting

https://www.koi.ai/blog/clawhavoc-341-malicious-clawedbot-skills-found-by-the-bot-they-were-targe...
1•Santas•20m ago•0 comments
Open in hackernews

TPP

https://github.com/universe-components/touchpoint
3•wchswchs•9mo ago

Comments

wchswchs•9mo ago
I've been exploring some new ideas while working on TPP, a protocol for multi-agent systems, and one of the concepts I'm really excited about is something I’m calling "Gear-Based Programming."

It’s a different take compared to traditional block-based programming—there are no explicit concepts like tasks or tools. The idea is to simplify the way we design and build intelligent agents, especially when it comes to multi-agent collaboration.

I think this could evolve into a new programming paradigm for the AI era—kind of like how object-oriented programming replaced procedural programming back in the day.

Curious to hear your thoughts. Does this idea resonate with anyone? Has anyone tried similar approaches?

wchswchs•9mo ago
In the past, building an adaptive system often meant defining dozens of different extension interfaces just to barely make it work. With TPP, that's no longer necessary—adaptive behavior can be achieved with just a single, unified type of interface.