8 Comments
User's avatar
Atharva Nihalani's avatar

incredible work- thank you both for writing this up

Jason Hausenloy's avatar

thank you for your terrific feedback!

Sophie Kim's avatar

yall COOKED

Noah Birnbaum's avatar

This was helpful.

One thing that I wonder (and maybe yall have answers to) is how much of the technical stuff in compute governance punt-able for when we get the political will/regime to have people seriously consider implementing them. What parts aren’t? What subset requires more time such that it makes it harder to punt?

Jason Hausenloy's avatar

a great question -- can imagine a version of this where we look at the calendar-time needed to implement the things

Canary Institute's avatar

I think if you're a technical person, it's worth helping on the issue now. The space over which politics can negotiate is strongly constrained by what is/n't technologically feasible. If you're a policy-but-not-tech person, then learning what is on/off the table politically and sharing with the techies is also valuable.

Sheikh Abdur Raheem Ali's avatar

Great post, saved.

I was playing around with Inkling NVFP4 earlier today (mostly on some no-CoT time horizon evals). Inkling includes an MTP head for speculative decoding, so I was planning to take a stab at extending Token-DiFR and Activation-DiFR from Luke's paper to it. I am using Modal's sandboxed runtime, which works well for me, and doesn't currently expose DCGM counters, but it's possible that may change in the future.