frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Show HN: Open-source engine running Gemma 4 26B in 2 GB RAM on any M-series Mac

https://github.com/drumih/turbo-fieldfare
675•gitpusher42•12h ago•236 comments

Show HN: CheapFoodMap – A map of good meals under $10

https://cheapfoodmap.com/
141•jaep1•10h ago•168 comments

Show HN: A local merge queue for parallel Claude Code agents

https://github.com/funador/claude-code-merge-queue
17•funador•3h ago•6 comments

Show HN: Qwen Scribe – local transcription and dictation for Apple Silicon

https://github.com/VladUZH/qwen-scribe
78•sidclaw•12h ago•17 comments

Show HN: Legionlinuxtui – Control Lenovo legion laptops in the terminal

https://github.com/nooneknowspeter/legionlinuxtui
3•nooneknowspeter•1h ago•0 comments

Show HN: Full stack development from one codebase web, Android, and iOS

https://github.com/usedowe/dowe-lang
4•victorriba•2h ago•1 comments

Show HN: GlobeHoppr – find a route between cities using limited-distance jumps

https://globehoppr.com
4•JohnJohnson123•1h ago•0 comments

Show HN: Kedge – Full-stack cloud with forkable VM snapshots and global SQLite

https://kedge.dev/
55•wgjordan•11h ago•15 comments

Show HN: RunNburn – Run a 295B Moe from a 98GB GGUF on a 64GB RAM Desktop

https://github.com/coderredlab/runNburn
2•coderredlab•1h ago•0 comments

Show HN: Damn Center (the page) – extension that helps centering any website

https://github.com/rinn7e/damn-center-extension
3•rinn7e•2h ago•1 comments

Show HN: A travel app not building generic itinerary

https://uncoverroads.com/
2•sourav_biswas•2h ago•0 comments

Show HN: OpenMetaHarness - complete long horizon tasks with more autonomy

https://github.com/openmetaharness/openmetaharness
4•IzhaarTejani•2h ago•1 comments

Show HN: IncomeOS – income from many small sources, in one screen and an MCP

https://github.com/Perufitlife/incomeos
2•renzomad•2h ago•0 comments

Show HN: The Federalist Papers, typeset as the 1787 newspapers they ran in

https://federalistreader.org/
37•vhwalke•13h ago•9 comments

Show HN: Dreeve, a self-hosted dashboard for your sports and fitness data

https://github.com/dreeveapp/dreeve
11•frogfuhrer•9h ago•1 comments

Show HN: Mousecrack – Bypass bot detection with deep learning

https://github.com/puffinsoft/mousecrack
7•BeverlyHills001•3h ago•0 comments

Show HN: Generate, Version and Host your resumé all via GitHub

https://github.com/sadigaxund/git-resume
2•sakhund•3h ago•0 comments

Show HN: I was tired of opening 2 tabs for every HN link, so I made a userscript

https://github.com/twalichiewicz/HNewhere
409•twalichiewicz•1d ago•114 comments

Show HN: Bullshit Detector – agent skills that fact-check videos and articles

https://github.com/SerhiiKorniienko/bullshit-detector
59•skorniienko•14h ago•63 comments

Show HN: XY – A Fast, composable, GPU-accelerated interactive plotting library

https://github.com/reflex-dev/xy
130•apetuskey•1d ago•46 comments

Show HN: An AI Color Grader – Pico

https://www.perbhat.com/pico
4•perbhat•4h ago•0 comments

Show HN: Always on visual context for your agents

https://useglass.up.railway.app/#
4•quantbagel•4h ago•1 comments

Show HN: Echologue – the private AI voice journal I built for myself

https://echologue.com/
29•arisAlexis•14h ago•13 comments

Show HN: Replicant Space – an HTTP API-based game based on the Bobiverse books

https://replicant.space/
7•mercutio•5h ago•0 comments

Show HN: Data structures as containers – sample chapter from my Big O book [pdf]

https://thebigobook.com/downloads/data-structures-sample-chapter.pdf
2•rcavezza•5h ago•0 comments

Show HN: AI Security Leaderboard – comparing cyber and CBRN safeguards

https://leaderboard.far.ai/
4•AdamGleave•5h ago•0 comments

Show HN: Play ROMs inline in your terminal

https://github.com/jhickner/rom
3•jhickner•5h ago•0 comments

Show HN: Dev-like – turn public engineering practice into agent skills

https://mrbro.dev/dev-like/
2•MarcusRBrown•6h ago•0 comments

Show HN: Email domain health check with shareable reports

https://shipmail.to/tools/email-health
2•jcoulaud•6h ago•0 comments

Show HN: Formally verified 3D CSG: Trust 93 lines spec, not 1000 lines AI code

https://github.com/schildep/verified-3d-mesh-intersection
113•permute•1d ago•48 comments
Open in hackernews

Show HN: RunNburn – Run a 295B Moe from a 98GB GGUF on a 64GB RAM Desktop

https://github.com/coderredlab/runNburn
2•coderredlab•1h ago
runNburn is an Apache-2.0 Rust inference engine for quantized GGUF models that are too big for your fast memory.

The core idea: weights stay file-backed (mmap), host residency stays under an explicit byte budget (--ram-budget), and GPU caches are sized from detected free/total VRAM — never from device-name presets. There is no conversion step, no sidecar cache files, no silent requantization. The GGUF on disk is the single source of truth.

The result that made me want to post this: Tencent's Hy3 (295B total / 21B active sparse MoE, a single 97.8 GiB Q2_K GGUF) runs on my desktop with 64 GB of RAM and one consumer NVIDIA GPU. The file is larger than RAM and VRAM combined; the selected experts for each token are pulled on demand (the newest path batches O_DIRECT reads through io_uring), while the pretrained routing is left untouched. On the same machine, same prompt, same decode length, a warm-run median gave ~5.5 tok/s decode vs ~2.0 tok/s for llama.cpp.

To be upfront about scope: for models that fit comfortably in VRAM, llama.cpp is still faster than runNburn today — its CUDA kernels have years of tuning and we measure against it honestly (interleaved A/B runs, medians, and any "speedup" that changes output quality is rejected). runNburn's lane is the model that doesn't fit.

What's in the box:

- CLI, interactive chat, and an OpenAI-compatible server (chat/completions + responses + conversations, SSE streaming, stateful continuation with KV/SSM snapshot reuse). It's built as a single-owner personal server — one active generation is the optimization unit; continuous batching and multi-tenant throughput are explicit non-goals. - Architecture-aware paths: Llama family, Phi, Gemma, Qwen dense/hybrid/MoE (including GatedDeltaNet layers), Nemotron-H MoE, Hy3, GLM — plus in-model multi-token prediction (self-speculative decoding) with device-side verification where the GGUF ships a drafter. - Backends: CPU is the default (x86 AVX2, ARM NEON), CUDA and Metal are active, Vulkan/OpenCL are experimental. Android works through a small C ABI (rnb.h). - Native quantized kernels for Q2_K–Q6_K, Q4_0, Q8_0 — including the low-bit K-quants that big-MoE builds actually ship in.

It's pre-1.0 and rough in places; recognition of an architecture doesn't mean every community variant works. But if you've got a model file bigger than your machine and you'd rather it run slowly than not at all, that's exactly the case it was built for.

Happy to answer questions about the offloading design, the expert-streaming path, or the measurement protocol.