Alexandr Wang

Yeah, train them to – well, actually, originally, like, you could – like, let's say you took a photo of this bottle. A machine wouldn't even be able to tell you what's in the photo. It would just know what the pixels were. But it wouldn't be able to tell you, like, oh, there's a bottle in that photo.

Yeah, train them to – well, actually, originally, like, you could – like, let's say you took a photo of this bottle. A machine wouldn't even be able to tell you what's in the photo. It would just know what the pixels were. But it wouldn't be able to tell you, like, oh, there's a bottle in that photo.

Yeah, train them to – well, actually, originally, like, you could – like, let's say you took a photo of this bottle. A machine wouldn't even be able to tell you what's in the photo. It would just know what the pixels were. But it wouldn't be able to tell you, like, oh, there's a bottle in that photo.

So, you know, one of the first AI breakthroughs was when YouTube built an algorithm that could tell when there were cats in their videos. And that was, like – In 2012 or 2011, this was this mind-blowing breakthrough that you could use an algorithm to figure out when there was a cat inside a video. And so AI, it started pretty simply with just how do we replicate vision?

So, you know, one of the first AI breakthroughs was when YouTube built an algorithm that could tell when there were cats in their videos. And that was, like – In 2012 or 2011, this was this mind-blowing breakthrough that you could use an algorithm to figure out when there was a cat inside a video. And so AI, it started pretty simply with just how do we replicate vision?

So, you know, one of the first AI breakthroughs was when YouTube built an algorithm that could tell when there were cats in their videos. And that was, like – In 2012 or 2011, this was this mind-blowing breakthrough that you could use an algorithm to figure out when there was a cat inside a video. And so AI, it started pretty simply with just how do we replicate vision?

How do you replicate basically the fact that our eyes and our brains can process all this imagery coming in? And that really led to, I think, one of the first major use cases of AI, which is self-driving cars. So this was – when I started the company in 2016, self-driving cars were all the rage because it was – there were all these like skunk works projects.

How do you replicate basically the fact that our eyes and our brains can process all this imagery coming in? And that really led to, I think, one of the first major use cases of AI, which is self-driving cars. So this was – when I started the company in 2016, self-driving cars were all the rage because it was – there were all these like skunk works projects.

How do you replicate basically the fact that our eyes and our brains can process all this imagery coming in? And that really led to, I think, one of the first major use cases of AI, which is self-driving cars. So this was – when I started the company in 2016, self-driving cars were all the rage because it was – there were all these like skunk works projects.

Yeah. Yep. That was all the rage. And so it was it was all about can you know, can we start building algorithms that can drive a car like a human would and do it safely and do it more efficiently? And that way, instead of that was that was one of the first major areas. OK. And then now, you know, I.

Yeah. Yep. That was all the rage. And so it was it was all about can you know, can we start building algorithms that can drive a car like a human would and do it safely and do it more efficiently? And that way, instead of that was that was one of the first major areas. OK. And then now, you know, I.

Yeah. Yep. That was all the rage. And so it was it was all about can you know, can we start building algorithms that can drive a car like a human would and do it safely and do it more efficiently? And that way, instead of that was that was one of the first major areas. OK. And then now, you know, I.

The whole industry has moved so fast, but then all of a sudden we got ChatGPT and we got more advanced stuff more recently that is able to talk like a human or sort of think like a human. And so it's really come pretty far recently, but all of it is about how do you build algorithms, how to use machines to be able to think like a person.

The whole industry has moved so fast, but then all of a sudden we got ChatGPT and we got more advanced stuff more recently that is able to talk like a human or sort of think like a human. And so it's really come pretty far recently, but all of it is about how do you build algorithms, how to use machines to be able to think like a person.

The whole industry has moved so fast, but then all of a sudden we got ChatGPT and we got more advanced stuff more recently that is able to talk like a human or sort of think like a human. And so it's really come pretty far recently, but all of it is about how do you build algorithms, how to use machines to be able to think like a person.

Yeah. So there's two parts. There's two parts to it. So the first part is you need really advanced chips. So you need like these, they're called GPUs or sometimes called TPUs or, you know, there's a lot of different words for it, but you need like the most advanced computer chips in the world. And they- How big is each one? Do you know?

Yeah. So there's two parts. There's two parts to it. So the first part is you need really advanced chips. So you need like these, they're called GPUs or sometimes called TPUs or, you know, there's a lot of different words for it, but you need like the most advanced computer chips in the world. And they- How big is each one? Do you know?

Yeah. So there's two parts. There's two parts to it. So the first part is you need really advanced chips. So you need like these, they're called GPUs or sometimes called TPUs or, you know, there's a lot of different words for it, but you need like the most advanced computer chips in the world. And they- How big is each one? Do you know?

They – I mean the biggest ones are actually like – the whole chips are like pretty big. But they're like a wafer kind of thing. But then you put a lot of them all together. So these – yeah. Yeah, yeah, exactly. These like – these chips. And the biggest ones are – yeah, exactly.

They – I mean the biggest ones are actually like – the whole chips are like pretty big. But they're like a wafer kind of thing. But then you put a lot of them all together. So these – yeah. Yeah, yeah, exactly. These like – these chips. And the biggest ones are – yeah, exactly.