Logan Kilpatrick

And we have a...

a state-of-the-art model uh for text-to-speech and it's it's really cool to see there's been a huge amount of demand um and wasn't the use case that i thought would be like particularly um would like grow particularly quickly but it actually is which has been really interesting to see there's like a huge amount of demand for like very native um audio sounding or like native human sounding uh text-to-speech video generation so you can go into ai studio and you can test this capability out and things like um

If folks have used Notebook LM, you could do these like podcast-like generations, which is really cool.

And there's lots of other use cases like that.

Very cool.

So you can make your own AI podcast in AI Studio?

Exactly.

You could make your own AI podcast.

In your voice.

Yeah.

In your own voice, which is awesome.

Wow, good to know.

This is part of the, like, there's actually not a big gap between being able to do that and what Notebook LM is doing, which is interesting.

Like, what I think people fail to realize about the Notebook LM story is that really the...

I think the audio piece was important and that's why we've externalized the audio piece with, with what we have right now, um, through the Gemini API and it's like our, our text to speech models.

But actually the thing that people were most compelled by was like the humanness of the scripts, um, in notebook LM and the human scripts were just written by Gemini.

There's no, there's no meds, like literally prompting it's prompt engineering and all that stuff.

Like there was no like secret sauce of what was happening.

Nothing.

It is.