logoalt Hacker News

spwa4yesterday at 9:14 PM0 repliesview on HN

These are native speech-to-speech models, so I think they've decided not to do that anymore.