frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: NanoWakeWord – Open-source wake word training for any device

https://github.com/arcosoph/nanowakeword
1•arcosoph_ai•2h ago
Hacker News,

Training custom wake words like "Hey Alexa" is often a resource-intensive task, demanding powerful hardware and complex manual tuning.

NanoWakeWord is an open-source framework designed to solve this. It features an intelligent engine that automates the ML pipeline, making it possible to build high-performance, production-ready wake word models with minimal effort.

What makes it different:

Train Anywhere, On Anything: The core architecture is built for extreme efficiency. You can train on massive, terabyte-scale datasets using a standard laptop or even a low-spec machine, all without needing a GPU. This is achieved through memory-mapped files that stream data directly from disk, eliminating RAM limitations.

Intelligent Automation: The framework analyzes your data to automatically configure an optimal model architecture, learning schedule, and training parameters. It removes the guesswork from building a robust model.

Total Flexibility and Control: While it automates everything, it also offers deep customization. You can choose from 11+ built-in architectures (from lightweight DNNs to SOTA Conformers) or easily extend the framework to add your own custom architecture. Every single parameter generated by the engine can be manually overridden for full control.

Smarter Data Processing: It moves beyond generic negatives. The system performs phonetic analysis on your wake word to synthesize acoustically confusing counter-examples, which drastically reduces the false positive rate in real-world use.

Ready for the Edge: Models are exported to the standard ONNX format. The framework also includes a lightweight, stateful streaming inference engine designed for low-latency performance on devices like the Raspberry Pi.

Try It in Your Browser (No Install Needed):

This single Google Colab notebook is a playground to train your first model. Inside, you can select and experiment with any of the available architectures with just a few clicks.

Launch the Training Notebook: https://colab.research.google.com/github/arcosoph/nanowakewo...

The goal is to produce models with an extremely low false positive rate (tests show less than one false activation every 16-28 hours on average).

The project is actively developed by Arcosoph, and all feedback or questions are highly welcome!

Key Links:

GitHub Repo: https://github.com/arcosoph/nanowakeword

PyPI Package: https://pypi.org/project/nanowakeword/

Pre-trained Models: https://huggingface.co/arcosoph/nanowakeword-models#pre-trai...

Discord Community: https://discord.gg/rYfShVvacB

Free website privacy and exposure scanner (no signup, instant results)

https://fshot.org/utils/surfacescan.php
1•victorkulla•46s ago•0 comments

Show HN: MemPalace Agent that sits in front of any LLM endpoint and gives memory

https://github.com/skorotkiewicz/mempalace/tree/agent/agent
1•modinfo•48s ago•0 comments

Rationale for the design of the Ada programming language

https://dl.acm.org/doi/10.1145/956653.956654
2•fanf2•1m ago•0 comments

NIST Weighs in on the Mystery of the Gravitational Constant

https://www.nist.gov/news-events/news/2026/04/nist-weighs-mystery-gravitational-constant
1•01-_-•1m ago•0 comments

Learn How to Structure a Test

https://andros.dev/blog/d6eb1348/learn-how-to-structure-a-test/
1•ibobev•2m ago•0 comments

AI is an extinction-level event for your rules of thumb

https://shippingbytes.com/2026/04/15/ai-is-an-extinction-level-event-for-your-rules-of-thumb/
1•gianarb•2m ago•0 comments

More tools for testing SQL dialects

https://buttondown.com/jaffray/archive/more-tools-for-testing-sql-dialects/
1•ibobev•2m ago•0 comments

Syntonic Dentiforms Redux

https://aras-p.info/blog/2026/04/13/Syntonic-Dentiforms-redux/
1•ibobev•3m ago•0 comments

Anthropic rolls out Claude Opus 4.7, an AI model that is less risky than Mythos

https://www.cnbc.com/2026/04/16/anthropic-claude-opus-4-7-model-mythos.html
1•01-_-•4m ago•0 comments

Show HN: Modulo – A daily spatial reasoning puzzle

https://www.sansparsh.art/modulo
1•Sansparsh•5m ago•0 comments

Insurance carriers back away from covering AI outputs

https://www.csoonline.com/article/4159292/insurance-carriers-quietly-back-away-from-covering-ai-o...
2•WaitWaitWha•5m ago•0 comments

Qwen3.6-35B-A3B on my laptop drew me a better pelican than Claude Opus 4.7

https://simonwillison.net/2026/Apr/16/qwen-beats-opus/
1•simonw•6m ago•0 comments

Show HN: Chitragupta Kafka Identity and topic level cost attribution

https://github.com/waliaabhishek/chitragupta
1•Pancake9738•7m ago•0 comments

India's TCS to probe sexual assault, religious conversion in office

https://www.reuters.com/sustainability/indias-tcs-probe-sexual-assault-religious-conversion-alleg...
1•satyapr93•7m ago•0 comments

Show HN: Stage – Putting humans back in control of code review

https://stagereview.app/
3•charleslpan•7m ago•0 comments

Ask HN: Is Opus 4.7 obsessed with malware for anybody else?

1•abstrct•8m ago•1 comments

Meta Platforms, Broadcom Partners to Co-Develop Multi-Gen Silicon AI Chips

https://techgraph.co/ai/meta-platforms-broadcom-partners-to-co-develop-silicon-ai-chips/
1•visitednews•8m ago•0 comments

Bootstrapping AI Evals from Context (Why 'Just Asking Claude' Fails)

https://scorable.ai/post/bootstrapping-ai-evals-from-context
1•Arimbr•10m ago•0 comments

Shader Lab, like Photoshop but for shaders

https://eng.basement.studio/tools/shader-lab
1•ragojose•11m ago•0 comments

Linux in Postgres in Linux in Postgres in Linux in Postgres

https://github.com/pglinuxiguess/pglinux
1•chaps•12m ago•0 comments

Tape 05 – Boards of Canada [video]

https://www.youtube.com/watch?v=6bghDcbzfEU
2•boarsofcanada•13m ago•1 comments

Why Gen Zers are trashing smartphones: 'People are just sick of it'

https://nypost.com/2026/04/16/tech/gen-z-and-parents-are-hitting-rewind-on-tech-people-are-just-s...
1•1vuio0pswjnm7•15m ago•1 comments

ISMS CORE – Self-hosted ISO 27001 GRC platform (Docker Compose, 23 frameworks)

https://github.com/isms-core-project/isms-core-platform
1•isms-core-adm•16m ago•0 comments

Snap to Cut 16% of Workforce as It Seeks Profitability

https://www.wsj.com/business/snap-to-cut-16-of-workforce-as-it-seeks-profitability-4030b891
1•1vuio0pswjnm7•16m ago•0 comments

Posting on the Internet Frightens Me

https://adayinthelifeof.nl/2026/04/16/posting-on-the-internet.html
2•jaytaph•22m ago•2 comments

Dwarkesh Patel's Podcast with Nvidia CEO Jensen Huang

https://thezvi.substack.com/p/on-dwarkesh-patels-podcast-with-nvidia
2•7777777phil•22m ago•0 comments

Show HN: Xata, open-source Postgres platform with copy-on-write branches

https://github.com/xataio/xata
2•tudorg•23m ago•1 comments

Gemini can now create personalized AI images by digging around in Google Photos

https://arstechnica.com/ai/2026/04/gemini-can-now-create-personalized-ai-images-by-digging-around...
1•gpi•24m ago•0 comments

Claude Code injects hidden prompts into file reads to stop malware tweaks

https://twitter.com/adrian_cooney/status/2044827025379123597
3•adriancooney•25m ago•0 comments

Many-Step Sequences in Go

https://lesiw.dev/go/step/
1•zimpenfish•27m ago•0 comments