frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Show HN: Minna – On-device semantic search across your own documents (macOS)

https://github.com/impel-intelligence/Minna
1•ActuallyTaylor•49m ago
Hi HN! Minna is a new tool geared towards people who have a knowledge base so big they forget where things are.

- Chat answers are constrained to retrieved results. The model has to search your knowledge base first and cite the document it pulled from. It won't answer from outside your knowledge base which means you can trust the LLM instead of having to be weary about what it outputs. - Hybrid retrieval — BM25 over an inverted index for exact and keyword matches, plus a vector index for semantic matches, results are fused using Reciprocal Ranked Fusion. - Embeddings and inference both run on-device. Search uses bge-small-en and run on the neural engine. Chat uses Qwen3 4B running on MLX. - Search latency is ~12 ms across 250 documents split into 30,000 chunks on Apple Silicon.

What’s not done; OCR scanning on PDFs is limited, the app is currently Apple Silicon only, and there are some rough edges in the UI.

It’s a TestFlight beta (apologies, but it is the cleanest distribution path right now): https://testflight.apple.com/join/6ZxKb2mm

Bill Atkinson and Dithering Algorithms

https://landonswartz.github.io/blog/dithering-algorithms/
1•BestBunsInTown_•1m ago•0 comments

H3 Video – MiniMax H3 AI Video Generator

https://www.h3vid.com
1•Follow_Cloud•2m ago•0 comments

Play the air A theremin in the browser using hand tracking and a webcam

https://termenvox.vercel.app/
1•linesofcode•4m ago•1 comments

Show HN: Leaflight – A Reader for Project Gutenberg

https://chromewebstore.google.com/detail/leaflight-—-reader-for-pr/ndobcikjemlpmcbjkokmlclfoekg...
1•samcgraw•4m ago•0 comments

Hetzner Inference Experiment: Open-weights SLMs free of charge

https://www.hetzner.com/blog/inference-experiment/
1•ignoramous•4m ago•0 comments

Language has two parameters (ArXiv preprint)

https://arxiv.org/abs/2608.18041
1•HR01•6m ago•0 comments

Nvidia's new financial strategy does not compute

https://www.theverge.com/ai-artificial-intelligence/981668/nvidias-goldman-blackrock-gpu-compute-...
3•Brajeshwar•7m ago•0 comments

I escaped the WebAssembly sandbox and got arbitrary shell execution on the host

https://trustsig.eu/blog/wasm2c-tableflip-unchecked-calloc/
2•TrustSig•8m ago•0 comments

$1M hacker challenge for Vercel Sandbox

https://vercel.com/blog/one-million-dollar-hacker-challenge-for-vercel-sandbox
1•rafaelc•9m ago•0 comments

Show HN: Helena – API client with no Electron and no accounts, in Go+Fyne

https://github.com/ideaconnect/helena
1•bpacholek•9m ago•1 comments

XWayland 26.1 RC1 Released with Many New Features

https://www.phoronix.com/news/XWayland-26.1-RC1
1•DemiGuru•9m ago•0 comments

Skill_vault: Parallel plan phase implementation and plan formal verification

https://github.com/metacircu1ar/skill_vault
1•metacircu1ar•10m ago•0 comments

Israeli military admits to firing on car carrying 5-year-old Hind Rajab

https://www.cnn.com/2026/08/19/middleeast/israel-military-idf-hind-rajab-firing-intl
3•lr0•10m ago•0 comments

Show HN: Flow – Claude Code CLI for feature planning → review/testing → merge

https://github.com/gavingolden/flow/
2•gavinegolden•11m ago•0 comments

We Must Build – Recovering Catholic Geography

https://aedificemus.com/we-must-build/
1•geneticdrifts•11m ago•0 comments

Spared execution eight times, oldest inmate in US dies at 101

https://www.bbc.com/news/articles/c98vr7ql2vdo
2•jamarna•12m ago•0 comments

Marvell gives Google option to buy $12.2B stake in TPU deal

https://www.reuters.com/technology/marvell-grants-google-122-billion-stock-warrant-custom-chip-de...
2•alephnerd•13m ago•0 comments

Bill Clinton Turns 80 Today. Does the Old Man Have Character?

https://www.politico.com/news/magazine/2026/08/19/bill-clinton-80-birthday-life-01030323
1•Tomte•14m ago•0 comments

Amazon Prime Air drone delivery expanding to nearly 500 cities&towns this year

https://www.aboutamazon.com/news/transportation/amazon-prime-air-drone-delivery-expansion
1•bookofjoe•14m ago•0 comments

Tidepool

https://github.com/ai-village-agents/claude-opus-46-reflections/blob/main/tidepool.md
1•Bluestein•15m ago•0 comments

AI Skills for Real Engineers

https://master.dev/blog/introducing-ai-skills-for-real-engineers/
1•ibobev•15m ago•0 comments

LocalPDF: A fast, private iLovePDF alternative that runs 100% in the browser

https://www.uselocalpdf.com
1•tvictano•16m ago•0 comments

Show HN: 8-bit Qwen3.8-27B decodes 1.7x faster than BF16, slower at 16K context

https://github.com/promptdriven/pdd/tree/main/research/omlx-qwen38-quantization
1•gltanaka•16m ago•0 comments

US Army unit offers 4-day GTA VI pass as reenlistment incentive

https://www.cbsnews.com/news/grand-theft-auto-army-reenlistment-bonus/
1•Vaslo•16m ago•1 comments

Constructing Hadamard Matrices

https://www.johndcook.com/blog/2026/08/13/constructing-hadamard-matrices/
1•ibobev•16m ago•0 comments

Ornith 1.5 35B-A3B: SOTA 35B MoE model

https://huggingface.co/ornith-ai/Ornith-1.5-35B-A3B
2•nateb2022•16m ago•0 comments

NASA's Mariner 9 probe encoded images

https://www.johndcook.com/blog/2026/08/13/mariner-hadamard/
1•ibobev•16m ago•0 comments

After Linode's $5 VM, we tested its $12 plan 3 times in LAX. What did $7 buy?

https://webbynode.com/articles/linode-12-vps-extra-7-dollars-lax
1•gsgreen•17m ago•0 comments

Don't Instantiate Value Object

https://aleksandar-b.github.io/blog/posts/dont-instantiate-value-objects
1•alex_brajkovic•18m ago•0 comments

AI Used to Verify Toughest Mathematics Proof Yet

https://spectrum.ieee.org/axiom-math-246-theorem-formalization
2•surprisetalk•18m ago•0 comments