Advanced Micro Devices Inc (AMD)

[CombJelly]

"Alan's take doesn't jibe with the article, which makes multiple consistent claims regarding performance boost,"

Maybe. All I have seen is that single line. Without examining the source, I don't know. Extraordinary claims need extraordinary evidence. And for CMOS, this is an extraordinary claim. If true, then IBM/AMD has a jump on the rest of the industry that is literally unprecedented. Assuming there isn't some undesclosed downside. Ok, potential downsides include a more complex process that cuts into yield and cycle time. But those are minor in comparison to the benefits. Ok, when reading a paper, what is the biggest flag that should trigger your bullshit detector? When the authors can't come up with a case where their proposal has problems. That means they either don't really understand the problem or they are hand waving. There are always costs/benefits, the trick is to realize the cases where it doesn't matter as much. In my opinion, they shouldn't be published if they can't be honest. Reality is that I am unpublished. That will be changing soon...