How sure are you that the next generation of video compression codecs won't rely in part on AI/ML models?
I'd even be surprised if a decade from now there are 0 models in the Linux kernel. If a SOTA model for I/O scheduling existed & was clearly superior for a broad range of scenarios, why wouldn't Linux adopt it as the default?
I'd even be surprised if a decade from now there are 0 models in the Linux kernel. If a SOTA model for I/O scheduling existed & was clearly superior for a broad range of scenarios, why wouldn't Linux adopt it as the default?