Corey Grant

And now with AI...

or data science, you're also using a GPU.

So it is kind of the base layer that allows, you know, parallel compute, leveraging CUDA, et cetera, to go out and do these, as I go, do-do-do-do, the combinations, permutations, all the data to make it work.

And NPU is very similar.

And those are made, for example, you know, there's Intel, there's Qualcomm.

Most of those are, there's not a, it's not a dedicated,

Okay, a GPU is dedicated VRAM that is separate from your system memory.

Yes.

An NPU generally, and I'm giving kind of broad because it kind of depends on the vendor, but an NPU is a separate device that kind of sets within the CPU architecture that will leverage your VRAM

to a certain extent to go do that AI processing.

So the real big difference between an NPU is think GPU is really for the massive compute and like the making of AI or an NPU is like, Hey, I have co-pilot and I don't want it to run slow.

I want an NPU or, Hey, I do small little things.

Hey, NPU.

So NPU GPU, most computers now on the market will come with an NPU.

It's almost like industry standard now, at least I think.

I mean, all of our new, you know, Dell Pro, all of them come with an MPU almost.

But not every computer comes with a GPU.

Yeah, absolutely.

I mean, it comes both fixed and mobile from the entry-level 500 series all the way up to the big daddy, the 6,000 Blackwell 96 gigs of RAM that will, like, literally shoot a rocket into space.

Okay, so let's talk about a tool that's completely transformed how I personally work.