What happens when you force 99/100 developers to write unit tests based on useless monster wrappers, mock classes around mock classes around mock classes around... ? What happens when you force 99/100 designers and architects to sleep by reading biblical and futile tech specs and requirements? Intel is sinking just as few many other fat realities. This is what happens when the management thinks that control is the best practice. Live long Test Driven Development, TDD is dead. Creation and genius is a spark in between the chaos.

Gomez Addams:

Here's my speculation : he left AMD for a BBD - bigger and better deal. He probably thought he had accomplished enough to put AMD on a course for success so his work there was done and that seems to have been proven correct. The bigger and better deal was a pile of stock options to do the same thing for Intel he did for AMD. It seems to me he is leaving Intel a bit prematurely because it remains to be seen what, if anything, he has contributed there. My guess, because none of my friends at Intel will say anything, is he did most of the architectural work he wanted to do at Intel and then he hit a wall of frustration. Frustration with Intel's inability get a working process at smaller dimensions than they have now. That is the key for his architectural work to be embodied in silicon because currently it can not be. They can simulate their processors all they want but until they are taped out and built in silicon it really doesn't matter. I bet Intel is sitting on a whole new generations of processor that currently can't see the light of day because their processes are unable to make them yet. Either that or they were so caught off-guard by AMD they bagged that generation and started over. That seems unlikely since JK has been at Intel long enough to have told them what to expect. Another guess is he got so frustrated with fighting to get things done they way he wants them done, which were different than how Intel had been doing things as you can see by their current designs as compared to AMD's, that he gave up and bailed. Either are fairly likely I think but we will probably never know unless he feels like telling us. My guess for the future is he will end up with a company designing ARM-based chips. There is a lot of activity with those for servers right now. Another possibility is Nvidia. He could go there and work on CPUs for the next generation(s) of DGX. Nvidia does work with ARM-based chips too so those two things might coincide.

1st assumption is definitely wrong. Transistor is transistor no matter if it is on 90nm or 7nm. And apparently, even if it had many more transistors per CPU core than current intel's offerings, they would just release products with fewer cores. (Still, if it had 30% higher performance per core. Nobody would care that it costs 30% more.) So, I doubt that he left because intel could not get 7nm working well enough. 2nd, there is chance. But I guess intel was willing to spend money to get adequate ROI. What would be possible is that while AMD's teams are much smaller, intel's may have had some extra friction as there are many more people playing corporate games.

This is same Guy that Designed Zen to the best of my understanding. Guy is very talented at what he does which means he can do Whatever the Hell he wants and when he wants, I like that attitude he brings iam the same way but in diffrent ocupations obviously.

he didn't design zen, he led the retraining and retooling of the engineers

One man seldom makes a huge difference, regardless of the lopsided amount of publicity given him. Intel has tens of thousands of employees. I do wonder why his tenure at Intel was so short--but I doubt it means much. Intel will sink or swim on its products--it's as simple as that (and so will AMD.)

If he were really making a huge difference he would never have left. It takes several years to build up a good team, build up momentum and only then do you go on a roll making amazing products. Once you have that you aren't going to let it go as you are now the A team changing the world. If he keeps doing 2 years and is off sounds like the A team never happened so he's off to have another go somewhere else.

Fox2232:

1st assumption is definitely wrong. Transistor is transistor no matter if it is on 90nm or 7nm. And apparently, even if it had many more transistors per CPU core than current intel's offerings, they would just release products with fewer cores. (Still, if it had 30% higher performance per core. Nobody would care that it costs 30% more.) So, I doubt that he left because intel could not get 7nm working well enough. 2nd, there is chance. But I guess intel was willing to spend money to get adequate ROI. What would be possible is that while AMD's teams are much smaller, intel's may have had some extra friction as there are many more people playing corporate games.

Obviously, opinions vary. It is my assumption that Intel has a design ready to go that is a leap forward in terms of IPC, core counts, and other factors, and their current issues with process advancement are preventing it from seeing the light of day. They are not talking about this because it could tank their stock value. Further, I believe they are not willing to redesign it for fewer cores on a larger process because of a variety of issues such as power consumption. Again, this is my opinion and it could very well be wrong. When Intel finally does get it together on their next generation process(es) it will be interesting to see what they announce as new products using it. Hopefully it is significant and competitive so prices drop and AMD is further inspired.