I know RAM is scarce and everything, but doubling down on LLM local acceleration with all of that dedicated silicon while at the same time sticking with Apple's traditional lack of RAM availability makes for a very weird product proposition to me.