Andrew is simply saying (if I read him correctly) that if you use generative models you might classify sounds into something other than Phonemes on one layer, a

Andrew is simply saying (if I read him correctly) that if you use generative models you might classify sounds into something other than Phonemes on one layer, and that classification may actually be better than "Phonemes" for recognising sounds. Phonemes are a construct for us to express a concept with language, but might not be the best way to represent a sound in a machine. So why try to write the routines to map sounds to phonemes when you can train it to just learn the categories that help it understand how to recognise the sounds better.