[moffkalast]

It's great that dolphins are getting audio decoders in language models first, does the Gemma team intend to roll that out for human speech at some point eventually too?