3 ms·Groq Inference Tokenomics: Speed, but at What Cost?34 points by hasheddan 3y agofragmede 3y agothats quite the parallelization! but if it gets us to a voice assistant that I can interrupt, I think they'll be making tons of money off it.