May 2, 2024 · 16 min · 8 segments
Mike and Amit talk about the giant chipmaker NVIDIA, and how it's creating the technology of tomorrow. \*\*\*\*\*\*\*\** Michael Wade is a Professor of Innovation and Strategy and Director of the…
If you look down right now in the machine that you're using, it runs obviously on a CPU.
Now, what's the difference? And I'm going to give you a layman's intuitive understanding, which is all I have.
By and large, CPUs are designed to do a small number of extremely massive calculations, right? So think of the really, really, really heavy kinds of lifting that a standard laptop or desktop or PC needs to do.
The idea is that CPUs, the better and better they get, the bigger and bigger calculations that they can handle.
Now think about your screen, right? Your screen's made up of thousands and thousands of pixels.
And every individual pixel has to turn on and off, or it has to show specific shades at certain points in time.
Now, the individual calculation required for each pixel is relatively simpler, but you've got to do thousands of these calculations in real-time simultaneously, and that's what GPUs were able to do.
Now, when we move to machine learning, Turns out that GPUs are much, much better at running things like deep learning algorithms because they also require thousands and thousands of relatively small calculations happening in parallel.
If you look down right now in the machine that you're using, it runs obviously on a CPU.
Now, what's the difference? And I'm going to give you a layman's intuitive understanding, which is all I have.
By and large, CPUs are designed to do a small number of extremely massive calculations, right? So think of the really, really, really heavy kinds of lifting that a standard laptop or desktop or PC needs to do.
The idea is that CPUs, the better and better they get, the bigger and bigger calculations that they can handle.
Now think about your screen, right? Your screen's made up of thousands and thousands of pixels.
And every individual pixel has to turn on and off, or it has to show specific shades at certain points in time.
Now, the individual calculation required for each pixel is relatively simpler, but you've got to do thousands of these calculations in real-time simultaneously, and that's what GPUs were able to do.
Now, when we move to machine learning, Turns out that GPUs are much, much better at running things like deep learning algorithms because they also require thousands and thousands of relatively small calculations happening in parallel.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.