Hacker News1d ago423 points94 commentslive
We turned EmbeddingGemma 2 into a speech-to-text model
EmbeddingGemma 2 is repurposed as a speech-to-text model that transcribes an hour of audio in ~13 seconds on a base MacBook Air using the Neural Engine, with a 5% word error rate.
Read the full story ateg2.haku.shWhy this is in the Signal
LAXIMA AI Signal curates the highest-velocity stories across Hacker News, GitHub trending, and new Hugging Face / Replicate model releases — quality-filtered, deduplicated, and refreshed every four hours. This item surfaced from Hacker News with 423 points (by amrrs). We link straight to the original source above — see the full live feed.
More AI Signal briefs
- HNICE Planning $100M 'Nightmare' Domestic Spy Program
- HNAI agents aren't 'going rogue.' Frontier labs are skipping security basics
- HNSmart people don't think others are stupid (2012)
- HNASML's strong orders ease fears AI spending boom is running out of steam
- HNSwitzerland proposes banning profiling-based ads to minors
- HNPasskeys Are Not a Trap
- HNAmazon pushes up prices across the Internet
- HNIosevka is an open-source, sans-serif and slab-serif, monospace font