Yeah, but the entire linked post is making the same kinds of assumptions about general trends in hardware. Now, those assumption might be stronger than these (I happen to think they are), but its logic really isn't terribly better than this.