... Hunters, Jo o here After being blown away by what @Sesame voice demo could do, I jumped at the chance to try their recently released open-source model (CSM-b1) in one of my apps. Bad news: it failed spectacularly - if anyone was successful applying it, let me know please. But ended up creating the Startup Roast Generator , with audio roasts powered by (with Audio -> from @ElevenLabs or @OpenAI TTS, sorry Sesame

... reservations. I was like OK, yea, I can see it's headed in the right direction, but would I WANT to talk to one over a human being, even if it meant no wait? Nope. Until today... @Sesame just launched their Conversational Speech Model (CSM) . I didn't expect Maya or Miles to be so natural. It's really sensitive so picks up a lot of external noise which kind of distracts them, but overall, their ability to mimic ... ... actually felt they were really listening to me and asking me follow up questions to get me talking more. This demo really makes me believe that the future where I choose to chat with an AI voice bot (vs