I'm an average user, not a chip architect, just someone who has had a passing interest in semiconductor design since I started building my own PCs as a teenager. In other words, I'm an average tech nerd, no more, no less.
I'm not old enough to be wise; I'm not young enough to know everything.
However, I've watched repeatedly as other tech enthusiasts, who are more or less at my experience level, worship Jim Keller as if he were the most brilliant man in the industry, the mythological Nichola Tesla of chip engineers. For instance, Moore's Law is Dead
released a video with the title "Jim Keller's Royal Core designed to Kill Zen 5". It has this as the thumbnail, despite Keller not being personally involved in the video in any way:
View attachment 15172
In recent years, I've seen this repeated cult hero worship of Jim Keller from within PC circles, and have never seen another engineer placed on a pedestal of honor as Keller has. For Intel fanboys, he's the savior who architected the phoenix with which Intel will arise from the ashes, reborn into its rightful role as the premiere technology company. This is the same genius savant who
got evicted from Intel for "personal reasons", and the circumstances of his departure are still unclear.
The overall
PC Master Race sees Keller's time at Intel as the kickstart of x86 designs finally overtaking Apple, leaving Arm designs in the dust, handily taking the efficiency and performance crown within the next couple of years. This miraculous new "Royal Core" project at Intel that Keller had spearheaded, along with the brain drain that the Nuvia exodus caused, has resulted in Apple's ultimate doom. The Mac has already fallen into the dustbin of history, a dead platform that simply hasn't realized that it has already reached the culmination point of its own demise.
As PC fanboys have predicted for decades, Michael Dell's advice for Steve Jobs in 1997 was right, and that Jobs should "shut it down and give the money back to the shareholders". Despite being the most valuable company in the world, and just three days ago
deemed the most valuable brand in the world, it's time for Tim Cook to close up shop, and finally be done with it. Only then will PC fanboys, both at MacRumors and beyond, be able to take a well-deserved respite from their holy quest.
Some of this mythos surrounding Keller probably has to do with him being relatively media savvy, at least for an engineer. He's been on Lex Fridman's podcast,
not just once, but
two times thus far. I watched part of them, and the only thing I can recall from either of those podcasts is that Jim Keller is related to Jordan Peterson. That's literally the only detail that I can remember from those interviews.
As I said, I'm by no means an expert in my own right. However, from the quotes that
@Cmaier has provided, even I know that Keller is barking down the wrong river. As I have learned from numerous intelligent individuals, including folks on this forum, the difference between RISC and CISC isn't trivial, and switching around a few gewgaws inside a chip with billions of transistors isn't going to magic a miraculous product into existence, one which can easily handle both x86 and Arm instructions with some translation circuitry tacked on.
In a previous thread I referred to Jim Keller as a "brilliant engineer" and
@Cmaier said that he wasn't sure if he was brilliant, but has worked with folks, many of which are not famous outside of select tech circles, whom could be classified as such. I was stating that mainly based upon Keller's surface reputation (and also just wanting to be polite) but am now rethinking much of what I had assumed about Keller. He is, at the very least, a smart man. Given what he has accomplished, he has to be intelligent.
However, I'm now left wondering how much is myth, and how much is reality. Nichola Tesla was undoubtedly brilliant, but much of his reputation today came about long after his death. I am left wondering how much of Keller's reputation is deserved, or simply based upon a
cult of personality.
AMD most likely made a mistake in not keeping an Arm project going, Intel had been given mana from heaven when
StrongARM fell into their laps after picking over DEC's desiccated bones, only to toss it all away when they
sold XScale to Marvell. (Not the comic book company, which would have had the same result.) Imagine where Intel would be now if they had kept an ARM v4 product that had been designed by the DEC Alpha team?
Intel's historical blunders aside, Keller's sentiment may be correct, but his statement boiling down modern CISC into nothing more than an internal RISC design shows that he's either playing it up for the cameras like an attention whore, or is making assumptions that are demonstrably untrue, as
@Cmaier and others have repeatedly explained with great detail.