William Hahn

speaker

749 appearances 2 recordings 1 series first heard Feb 2025 last heard Apr 2025

William Hahn’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

No recordings in the last 12 months.Older appearances are listed below; set an alert to hear about the next one.

Appearances

newest first · ▶ plays the moment

Theories of Everything with Curt Jaimungal · Barenholtz, Hahn, Jaimungal: AI Will Put 1000 Doctors Per Person · 7 Apr 2025 podcast

But what we gotta remember is while that model can't examine a face or listen to phone uh to to the voice changes.

There are other, what used to be called narrow AI models that have been developed for the last 10, 15 years, some longer, that can do that.

Now they're very cumbersome pieces of code.

They're not easy to work with, and you cannot talk to them in English.

But in the next couple of years, those two kind of fields are going to be merging, these traditional what deep learning type AI and the language model.

And then I think they're absolutely going to be able to look for face droops and voice changes and shuffle gate analysis and things like that.

I I I think it's debatable whether or not they can analyze emotion.

I I would argue that they could they can.

There's a new one that just came out.

I'm sure some of you have seen we once we went to yeah, we wanted to run it later um is this sesame, and it's incredibly good.

At not just recognize your emotion, but at predicting emotion.

We were just talking to it yesterday and we were saying, well, how is it because it claimed it didn't have emotions, but it emotes, right?

It's very expressive.

So in general, I think it's the safest strategy is even if you think it can't do that thing today, XYZ, assume it will be able to in the next six to twelve months.

We we also have different modalities with which we experience the world.

This sort of classic left versus right brain paradigm.

And one part of us sees the world in this sequential, uh serial kind of fashion, and the other part of our mind sees things all at once, like recognizing a face.

We don't really decompose that, it's sort of just there.

And um, I suspect that maybe these different mechanisms that were discovered.

Covering in technology, auto regression and diffusion, maybe these are both valid models.

Showing 81–100 of 749 · page 5 of 38