HackerTrans
TopNewTrendsCommentsPastAskShowJobs

gkapur

no profile record

Submissions

Surfacing a 60% performance bug in cuBLAS

kernelspace.substack.com
10 points·by gkapur·vor 2 Monaten·0 comments

A Principled ML Compiler Stack in 5k Lines of Python

kernelspace.substack.com
1 points·by gkapur·vor 2 Monaten·0 comments

Agentic systems redraw the Pareto frontier on ARC-AGI

poetiq.ai
7 points·by gkapur·vor 8 Monaten·1 comments

comments

gkapur
·vor 2 Monaten·discuss
On the limitation side:

Do you think this would scale to larger transformer models with more parameters per layer?

How would this work with MOE models or sparse models?
gkapur
·vor 9 Monaten·discuss
Adding to the prior comments as my intuition matched yours, there’s a nice Reddit thread that gives some context into how it can be faster even if you require exact matches: https://www.reddit.com/r/LocalLLaMA/s/ARxHLqRjdM

The TLDR/key (from my understanding) is that verifying N tokens can be faster than generating N tokens.