A modern self-referential weight matrix that learns to modify itself

Re:A modern self-referential weight matrix that learns to mo... by armchair_ on Wednesday April 13, 2022 @04:18PM (#31019735)

Well, the code that was published alongside the article is written in Python and CUDA, so you're not looking at the right kind of processor to start.

My 5-year-old, consumer-grade GPU does 1.5 MHz * 2300 cores, whereas the equivalent released this year does 1.7 Mhz * 8900 cores. Granted, not the best way to measure GPU performance, but it is roughly keeping pace with Moore's law, and it's going to be a better indicator of the future than Intel CPU capabilities, especially for machine learning applications.

[ Parent | on HN ]

Re:A modern self-referential weight matrix that learns to mo... by erdos4d on Wednesday April 13, 2022 @04:38PM (#31019936)

So you are saying that we can get another 3 decades of Moore's law by switching to GPUs made of silicon rather than CPUs made of silicon? Well fuck, problem solved then. I was completely unaware it was so easy.

[ Parent | on HN ]

Re:A modern self-referential weight matrix that learns to mo... by VyperCard on Wednesday April 13, 2022 @06:33PM (#31021139)

Yeah, but mostly for matrix multiplications

[ Parent | on HN ]

Re:A modern self-referential weight matrix that learns to mo... by ponow on Thursday April 14, 2022 @04:52AM (#31024635)

GHz not Mhz for your GPU specs.

[ Parent | on HN ]