Not yet. In a few years, consumer GPUs or dedicated ML units (which Apple SOCs have) and local storage and ram will be large enough to run “a few years old” LLM

Not yet. In a few years, consumer GPUs or dedicated ML units (which Apple SOCs have) and local storage and ram will be large enough to run “a few years old” LLMs. The cutting edge will always be in the cloud but a market for personal AI does exist.