[SMAAART]

Can someone help me make some sense of the 10 gigawatts of compute?

Colossus data has 230K GPUs (150,000 H100 GPUs, 50,000 H200 GPUs and 30,000 GB200 GPUs) [source https://x.ai/colossus]

Energy usage: up to 150 megawatts of electricity per day [source: https://en.wikipedia.org/wiki/Colossus_(supercomputer)]

So, when SamA talks about 10 gigawatts of compute does he mean per day or GWH (Gigawatts-hour)?

[Balgair]

Video. It's for video.

We don't have the compute to do video on demand right now like we do images or text or audio.

Combining all the modalities together, smoothly, at speed, and for cheap, is going to take a hell of a lot of thinking sand powered by magic rocks.

[rsfern]

I think it’s just scale to the moon rhetoric, like “what if we used 100x more compute?”. Since the units are power and not energy, I’m going with 10 GW continuous load (for training? inference?) but I think it’s not exactly meant literally