the localhost:0006 | how it works

You know, I go back to thinking about AI and how does it work.

And one of the things that I'm going to do is that we're going to throw up a bunch of slides while we're talking about how it works.

I like to think of AI as a stack and it kind of fits into that space.

And when we're thinking about that stack, we think of it as silicon being at the bottom.

Then we have this runtime layer that is this interface, almost like the operating system.

There's the model that goes on top of it.

There's context that we bring to it, like what do we bring to the equation? Like what type of questions are we asking? And then there's the interface that we use to interact with it called the application layer.

And so we're going to go over that.

Each one of us is going to take a little bit of that layer But I'm going to jump in first into the silicon layer.

Silicon to me is this idea that the hardware is the core basis of where everything operates.

And there's many different levels or layers in that silicon.

One of them is the CPU.

And you can think of the CPU as like the head chef.

They're the versatile ones.

They're good at communicating orders.

They're making decisions.

They're coordinating everything.

They can do almost any job, but they're kind of good at everything, but they're not experts at some specific things.

And so they're very well-tuned to take on multiple tasks and deal out things, but there are some specialists in that space.

If you brought in the GPU, the GPU would be like this huge team of prep cooks.

They can take a mountain of similar work and tackle it all at the same time.

That's great at drawing millions of pixels or turning out what is great math behind the AI.

And then there's the NPU.

This is kind of a special purpose.

It's kind of like a kitchen appliance.

You know, I go back to thinking about AI and how does it work.

And one of the things that I'm going to do is that we're going to throw up a bunch of slides while we're talking about how it works.

I like to think of AI as a stack and it kind of fits into that space.

And when we're thinking about that stack, we think of it as silicon being at the bottom.

Then we have this runtime layer that is this interface, almost like the operating system.

There's the model that goes on top of it.

There's context that we bring to it, like what do we bring to the equation? Like what type of questions are we asking? And then there's the interface that we use to interact with it called the application layer.

And so we're going to go over that.

Each one of us is going to take a little bit of that layer But I'm going to jump in first into the silicon layer.

Silicon to me is this idea that the hardware is the core basis of where everything operates.

And there's many different levels or layers in that silicon.

One of them is the CPU.

And you can think of the CPU as like the head chef.

They're the versatile ones.

They're good at communicating orders.

They're making decisions.

They're coordinating everything.

They can do almost any job, but they're kind of good at everything, but they're not experts at some specific things.

And so they're very well-tuned to take on multiple tasks and deal out things, but there are some specialists in that space.

If you brought in the GPU, the GPU would be like this huge team of prep cooks.

They can take a mountain of similar work and tackle it all at the same time.

That's great at drawing millions of pixels or turning out what is great math behind the AI.

And then there's the NPU.

This is kind of a special purpose.

It's kind of like a kitchen appliance.