Kieran Kunhya

Because cost is increasing, because we need more power.

And so you're limited by either your CPU, your RAM, or your networking.

And you need to optimize.

And this is where...

value is going to be especially because like doing ai is going to help do the programming of like business right and so the core thing that you will not be able to vibe code are optimization for the hardware to be as fast as is possible i'd love to talk to you about who and how should learn assembly but first i think we need a bathroom break

There is not enough data to train on.

And this is the biggest issue.

I started my career actually doing some assembly for titanium, right?

So the titanium is a dead processor type, right?

Which was done by Intel and HP a long time ago when they wanted to do 64 bits.

Well, they lost and then we got AMD who did it.

AMD 64, which became x64.

But Itanium was extremely interesting in the sense that those were processors who had a ton of computing power to do floats, FMAs, which is similar to what we need now for LLMs, right?

And you could pack three operations per line that could be loaded.

So basically, you had an output of basically 6 billion of operation per second.

But the bus, the memory bus only allowed 1.5, right?

So your CPU was four times faster.

So you had to do crazy things to pack things in memory, reuse the registered data.

And those type of semantics, no language could do that, right?

So, like, I have the Itanium programming book because Intel did amazing books, but that's exactly what Kiran says.