Mati Staniszewski

speaker

2,430 appearances 3 recordings 1 series first heard Sep 2026 last heard 6 Sep

Mati Staniszewski’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 3 in all, peaking in Sep 2026 with 3.

newest first · ▶ plays the moment

So you you you pronounce it in a very different way when you read it out loud.

Um so we did it with text-to-speech.

That was like the the big step change because previously all the voices were pretty robotic.

And then over the years, we've built all audio models.

Um and the reason we started with voice is we thought that for solving the communication, communication with AI, which we want to solve, the voice.

is the the biggest gap that exists today.

And it also will be the biggest opportunity because to access the benefits of AI you will need to learn how to communicate with it.

And we want to solve that.

Such a great question.

Um it doesn't have to, is the quick answer.

Today it usually does because you can control what's happening in the middle.

You can swap in an elements for the LM or translation or text generation.

So you can swap it in and out much easier.

And you have that monitoring of what happens in the time.

Usually it's more reliable today.

The text.

The having them as separate steps.

Um, but we are seeing that in certain cases.

You want to bring them closer together and fuse them.

So on the research side, effectively, you know, we start with text-to-speech, but then we added all audio models.

Showing 301–320 of 2,430 · page 16 of 122