Dwarkesh Patel
👤 Speaker
15267
total appearances
Appearances Over Time
Podcast Appearances
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
Hyperscalers are spending... This is shocking.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
Dylan said they're spending 50%.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
of their CapEx this year.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
On memory?
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
On memory.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
That's believable.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
What is hyperscaler CapEx?
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
It's like high hundreds of billions, maybe a trillion.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
And they're spending half of that on memory.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
Okay, so that is a huge constraint.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
That's why we're not going to get new laptops and phones this year.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
But at the same time, we have too much memory.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
People are willing to put too much memory into these systems.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
Right.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
Why is JetSense shoving all this memory into these racks?
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
if you don't need it.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
This is how big the train that takes up every 20 milliseconds is.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
Yeah, and sorry, a very basic question.
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
This is what is actually done?
Dwarkesh Podcast
Reiner Pope – The math behind how LLMs are trained and served
Mm-hmm.