Logan Kilpatrick

And it works.

It works pretty good.

So this is good.

Let's try it.

Party.

Okay, to be honest, whispering like that gives like nails on a chalkboard to me.

But you can give it the feedback.

yeah we can't change the prompt it said speak in a calm whisper uh which actually is a great example of like how good the tts model is our texas right the fact that it was able to like whisper like that um i just personally don't want the the whisper so yeah is that a relatively recent tts model or or this is the new one so we actually just upgraded this model i think like a week or two ago or something like that so it is um it's not yet on the three model

the three model family but it's like I'm pretty sure it's actually state-of-the-art the TTS model is right now which is really cool so you can you can try it in a studio you can build experiences with it in it um it has a lot of the upgrade was around like control and steerability um and the audio fidelity which is uh which is super cool to see awesome your uh your speech to speech seems to have taken a decent leap as well in recent months

More to come on speech-to-speech too.

We just rolled out last week or the week before the real-time translation capability as well, which is based on the same speech-to-speech stuff.

So we'll have that for developers early next year, which is really exciting.

So you'll be able to build like real-time speech-to-speech apps that like translate between different languages.

So I think that would be super cool to see people come up with new ideas there.

Agreed.

Yeah, I think it's a natural.

We're betting a whole lot on audio as this sort of like next form factor.

I think like audio and code together going into 2026, like there's something interesting and magical there.

So I'm really excited to see what happens.

what ends up happening.