Pat Gelsinger Is Going Back To Intel As New CEO

[bug77]

I saw him speak when I was in grad school way back in about 2004 and he gave a fascinating presentation on what we would today call IoT. Some of the concepts he was pushing were still well in advance of what you see commercially available right now, and remember that this was years before anyone had even heard of an iPhone.

This guy is technically smart and has a great deal of charisma, which are two major assets you need in a CEO. This is very good news for Intel.

Talking about something resembling IoT in 2004 does not make one a visionary. There was much talk back then about how our refrigerator will be able to tells us when we're about to run out of milk and things like that. The whole idea behind Java was to power these "IoT" devices. To the point Sun imagined processors capable or running Java natively: https://en.wikipedia.org/wiki/PicoJava

The guy is obviously smart and knowledgeable, but that is based on position he previously held, not on him knowing about IoT 15 years ago

I fear Bob Swan is too much the economics guy - not having a deep understanding of the technology.

The trick is, there are times when you need a tech guy in charge and there are times when you need a "bean counter" more. And it's really a trick, because companies never seem to make the change in a timely manner.

Example of tech guy holding the position for too long: AMD's Hector Ruiz.

[Paradigm Shifter]

I hope this doesn't mean Intel decide to drop oneAPI before it has a chance to grow. We really do need competitors for CUDA.

[chuckula]

Talking about something resembling IoT in 2004 does not make one a visionary. There was much talk back then about how our refrigerator will be able to tells us when we're about to run out of milk and things like that. The whole idea behind Java was to power these "IoT" devices. To the point Sun imagined processors capable or running Java natively: https://en.wikipedia.org/wiki/PicoJava

Yeah... you obviously weren't at the talk. Trust me, he would have been laughed out of the room if he said you could check email from a refrigerator. Try this: Scada microsensors embedded in building materials that harvest thermal energy and auto-establish a mesh network for communication. That wasn't "dur email in refrigerator" and even in 2020 you don't see that around every corner even if 16+ years of research have made it closer to feasible. There's a reason I remember the talk 16 years later, and a CPU that happens to have a architecture for running Java isn't anywhere near that interesting.

[horizonbrave]

Is Itanium like Titanium but without a T, I mean like a typo?

[Guest]

I sure hope Intel's open source efforts continue unabated.

[ThoreauHD]

Hard to be a futurist when the present is turning to ash, but I guess we'll see what he can do, again.

[bug77]

Yeah... you obviously weren't at the talk. Trust me, he would have been laughed out of the room if he said you could check email from a refrigerator. Try this: Scada microsensors embedded in building materials that harvest thermal energy and auto-establish a mesh network for communication. That wasn't "dur email in refrigerator" and even in 2020 you don't see that around every corner even if 16+ years of research have made it closer to feasible. There's a reason I remember the talk 16 years later, and a CPU that happens to have a architecture for running Java isn't anywhere near that interesting.

I was just saying, there was a lot of talk about interconnecting stuff even back then. In various forms.

[DarkCloud]

So much for new blood at Intel, going back to the 90's. Perhaps we will see another Itanium product or XScale dud from that era. The old stogy middle and upper management must be breathing a sigh of relief in knowing they are just rearranging the deck chairs.

[Slartifartblast]

Good luck Pat, you have your work cut out and possibly have a more difficult job than being Trump's defence lawyer right now.

Don't listen to activist investors and double down on making the leading edge process nodes work more than anything else because where the current ones are a strategic risk for supply.

[L_A_G]

On the whole I'm happy to see Swan, a finance guy poached from eBay, go, but considering where his replacement is coming (VMWare) from and what he's presided over at Intel in the past (Netburst, Itanium/"Itanic", etc.) I'm not entirely optimistic either. (The reason why I'm apprehensive of him coming from VMWare is due to having fought with their vcloud software at work and virtualbox being my "It's free and it works, somehow" local VM)

Another thing that worries me is if he actually outsources their foundries. Will they just do what AMD did with Globalfoundries and move their high end silicon to TSMC's already overstretched higher end nodes? Or will they continue to use their own foundries, just that those foundries are now more free to acquire non-Intel foundry customers (compared to today where their non-Intel customers are essentially trade secrets).