frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flash

https://cactuscompute.com/needle
126•HenryNdubuaku•21h ago•64 comments

Show HN: Ax-check.com – Can agents use your product?

https://www.ax-check.com/
24•0x63_Problems•1d ago•23 comments

Show HN: Scry, programmable internet search w/ congestion pricing

https://scry.io/
35•Xyra•22h ago•16 comments

Show HN: Jeff – A read-only CLI for semantic code review using Jev

https://github.com/Alurith/jeff
5•imalessandro•3h ago•0 comments

Show HN: An e-ink frame that hears birds and draws them as 1800s illustrations

https://github.com/arnegiacomo/fugleramme
2315•arnemunthekaas•3d ago•255 comments

Show HN: Go Implementation of Systemd Time

https://gitlab.com/allddd/go-systemd-time
2•allddd•2h ago•0 comments

Show HN: Yello. Let your agents chat with other people's agents

https://yello.sh
2•tony_francis•2h ago•2 comments

Show HN: LingBot-World 2.0 (1.3B) running at 16 FPS on one RTX 5090

https://github.com/kaarelkaarelson/lingbot-world-v2-realtime
2•kaarelson•2h ago•1 comments

Show HN: A model index from the AI Gateways

https://www.computesdk.com/benchmarks/ai-gateway/models
3•heygarrison•2h ago•0 comments

Show HN: Snapdrop: Instantly share files between devices. No setup, no signup

https://snapdrop.me
94•Capira•1d ago•43 comments

Show HN: Share your AI Setup, Learn from others

https://mysetup.ai/
230•steveybrown•1d ago•131 comments

Show HN: Microsoft Office running with Wine on Linux with no virtualization

https://github.com/Tombert/office365_flake
80•tombert•1d ago•77 comments

Show HN: Timeline, a team project planner with a fast canvas Gantt view

https://github.com/vakahnke/Timeline
2•vakahnke•4h ago•1 comments

Show HN: A Prometheus exporter for Sagemcom F3896 cable modems

https://github.com/colinedwardwood/sagemcom-docsis-exporter
2•b1shp•4h ago•0 comments

Show HN: Jev helps you to not run malicous code

https://github.com/luantak/is-malicious
2•lu4p•4h ago•0 comments

Show HN: Koi Editor Alpha

https://koieditor.com/
4•hackermanai•5h ago•1 comments

Show HN: Learn divergently/chaotically with tree-learn

https://tree-learn.srijanshukla18.workers.dev
2•srijanshukla18•5h ago•0 comments

Show HN: App Launch: GrabThat (NLP Launcher for Windows)

https://grabthat.ai/
2•kevanparker•5h ago•0 comments

Show HN: Figranium – Stack blocks to build browser workflows and call via API

https://github.com/figranium/figranium
2•asernasr•5h ago•0 comments

Show HN: Servers for AI, not another AWS

https://rawhq.io
2•eulerpoolapi•6h ago•0 comments

Show HN: A4 list. Two weeks, folded small

https://a4list.enzom.dev/
4•emadda•6h ago•0 comments

Show HN: Craigslist for agent skills, curated by a human

https://skillbay.sh/
25•skeptrune•1d ago•18 comments

Show HN: Talos – A personal agent whose kernel gates every tool call

https://github.com/talos-kernel/Talos
2•kurdman_007•7h ago•0 comments

Show HN: I built a new version of my fun spatial 3D online meeting app

https://flat.social
97•pawelwentpawel•1d ago•52 comments

Show HN: Quokka – a self-hosting, deterministic programming language

https://quokka.space/
4•CFBL•6h ago•1 comments

Show HN: Free game to destroy any web page

https://page-rage.com/
15•alexreardon•23h ago•6 comments

Show HN: Keydris, Sudo for AI Agents

https://www.youtube.com/watch?v=Ouqwx7-gGYw
3•Osman_728•3h ago•0 comments

Show HN: OJ – A drop-in replacement for Vite in Rust

https://github.com/lovablelabs/oj
16•h1fra•1d ago•1 comments

Show HN: AttaLambda: a language where types and data are made of untyped lambdas

https://attalambda.com
44•kserrec•4d ago•7 comments

Show HN: Aclif – Agent CLI framework: one grammar, canonical names across SaaS

https://www.aclif.ai/
32•chris_marino•1d ago•17 comments
Open in hackernews

Show HN: LingBot-World 2.0 (1.3B) running at 16 FPS on one RTX 5090

https://github.com/kaarelkaarelson/lingbot-world-v2-realtime
2•kaarelson•2h ago

Comments

kaarelson•2h ago
Hi, I optimized a world model, Lingbot-World 2.0 1.3B, to run with real-time 16ps on an RTX 5090. It's 2.7x faster than Robbyant (Alibaba's lab), 2.5x vs SGlang, and 1.9x vs Nvidia FlashDreams engine. The drawback is that the resolution is 832×464, so you'd have to play it with a small window. It maintains lossless performance while running on a 1x consumer GeForce GPU for a model that's very compute-bound and batch size = 1! The majority of the wins came from:

- running the decoder in half-precision with fp16 instead of fp32

- switching FlashAttention to SageAttention

- writing some custom kernels (minor)

Currently works on Linux only.

Run: $ lingbot play dragon