Steven Sinofsky discusses an Nvidia and Microsoft PC announcement and describes the shift toward graphics and neural processors for AI workloads.
He argues that running models locally could reduce recurring token chargesOn-device inference runs a trained model on the user's device rather than relying entirely on a remote model service. for some uses, while acknowledging uncertainty about hardware and software integration.
Sinofsky compares PC and Apple ecosystems and reflects on Surface and backward compatibility. His predictions about local AI, memory costs and future devices are presented as opinions.
Watch on YouTube
