Google's Gemma 4 open AI models use "speculative decoding" to get up to 3x faster

Ars Technica AI
Generative AI Open Source AI AI Research

Up to 3x the speed with no loss of quality - is it too good to be true?