Mozilla releases dataset and model to lower voice-recognition barriers

The browser maker has collected nearly 500 hours of speech to help voice-recognition projects get off the ground.