Article [NYT] Intel’s Culture Needed Fixing. Its C.E.O. Is Shaking Things Up.

[Hitman928]

Intel’s Culture Needed Fixing. Its C.E.O. Is Shaking Things Up. (Published 2020)

Robert Swan, who leads the world’s biggest chip maker, is pushing his 110,000 employees to confront internal problems more openly.

www.nytimes.com

Short article but has some interesting tidbits about the changes happening at intel.

 

[JDG1980]

It's easy to say you're changing the culture. It's a lot harder to actually do it. I guess we'll see.

 

[moinmoin]

The crises will be among “the best things that ever happened to the company,” Mr. Swan said

A little early to state that, but good luck. Intel certainly needs changes.

Intel recruited Mr. Keller, an engineer known for stints at Apple and an Intel rival, Advanced Micro Devices, in July 2018 to spearhead other design changes. He set up small teams, instead of the large ones that typically designed most of the chip circuitry from scratch to maximize computing performance. Some of the new teams were directed to develop standard blocks of circuitry that could be reused in different products so engineers could work independently without waiting for others to finish their portions of a chip.

That helped reduce the time needed to make and test chip design changes from weeks to days, Mr. Keller said. Creating an entire chip design can now be as much as three times faster, he said.

Pretty insane they didn't do that before already.

 

[Nothingness]

I know for sure that Intel had been reusing blocks for years (in particular datapaths). Perhaps Keller made that more widespread.

 

[DrMrLordX]

Is this more of a puff piece than an actual news story? It definitely seems to gloss over some of Intel's biggest problems with 10nm. Interesting that they went into a little of what Jim Keller is doing there, though.

 

[lobz]

“Where Bob is doing a great job is to give them some modesty.”

Certainly not publicly... Intel appears more cocky to me right now than ever, and the stunts Ryan Shrout pulls are nothing short of disgusting and truly pathetic.

 

[JasonLD]

A little early to state that, but good luck. Intel certainly needs changes.


Pretty insane they didn't do that before already.

Well, when it worked well, that approach used to be the Intel's biggest advantage over competitors.

 

[Asterox]

“Where Bob is doing a great job is to give them some modesty.”

Certainly not publicly... Intel appears more cocky to me right now than ever, and the stunts Ryan Shrout pulls are nothing short of disgusting and truly pathetic.

You can be cocky all day and night, but unless you have a good foundation being cocky is like morning fog.

Well Ryan is doing good job, his blupers are great they hit straight into the.............................

https://twitter.com/x/status/1234235424748515330

https://twitter.com/x/status/1234230582814572544

 

[exquisitechar]

Good luck to Mr. Swan, he's gonna need it to fix Intel's issues.

You can be cocky all day and night, but unless you have a good foundation being cocky is like morning fog.

Well Ryan is doing good job, his blupers are great they hit straight into the.............................

https://twitter.com/x/status/1234235424748515330

https://twitter.com/x/status/1234230582814572544

I'm sure Intel is panicking over Mindfactory sales and revenue, but their 10x higher than AMD's quarterly revenue should provide them some comfort!

 

[moinmoin]

Well, when it worked well, that approach used to be the Intel's biggest advantage over competitors.

Slow prototyping is an advantage?

 

[lobz]

Good luck to Mr. Swan, he's gonna need it to fix Intel's issues.

I'm sure Intel is panicking over Mindfactory sales and revenue, but their 10x higher than AMD's quarterly revenue should provide them some comfort!

Not all big OEM contracts last forever.

 

[JasonLD]

Slow prototyping is an advantage?

No, about having a large team designing most of the chip circuitry from scratch. That did enable the tighter integration of their nodes and design. Like I said, when it was working well, it was considered Intel's advantage over competitors, but doesn't give you much flexibility for plan B when something goes wrong, like during the times of Intel's 10nm troubles.
I am sure Keller's approach would be better for the designs using multiple chiplets using Foveros/EMIB.

 

[moinmoin]

No, about having a large team designing most of the chip circuitry from scratch. That did enable the tighter integration of their nodes and design. Like I said, when it was working well, it was considered Intel's advantage over competitors, but doesn't give you much flexibility for plan B when something goes wrong, like during the times of Intel's 10nm troubles.
I am sure Keller's approach would be better for the designs using multiple chiplets using Foveros/EMIB.

Even when it working well it's a waste of time and manpower to have it as the sole approach available. You could have both, small teams doing quick prototyping with modules, then large teams creating the chip circuitry basically from scratch based on the former's findings and decisions.

 

[krumme]

It's easy to say you're changing the culture. It's a lot harder to actually do it. I guess we'll see.

Yeaa. Good luck to him. Its a pain to do.
Takes lots of conflicts. Frustration and anger.
You take managers privileges. Get rid of people. Try to inject hope and energy.

Man. It takes many years.
I hope the best for them. We need a strong Intel.

 

[aigomorla]

Its not a culture thing its a price issue.
If they had kept prices streamline and not tried to stuff there wallets, they wouldn't have been in such a slump.

Also, they just don't care about listening to us.

Before they had a monopoly on the market, but with Ryzen, people are just fed up and now have a different direction to goto.
This is what broke the camel's back.
If they had these prices at Ryzen Launch, people wouldnt of been in mass Exodus to Camp AMD.

I was an avid pro Intel supporter, even had high branch internal contacts and had to sign NDA's with them, they would hand me cherry picked ES samples to post over at XS and here. I even got my samples before Anandtech did especially with the entire Nehalem arch.

But Intel is definitely not the same company i remember working with.
Anyhow, i really would like to see Intel clean up house and make stuff great again, but it sounds like a Trump campaign, only Make America Great is replaced by Make Intel Great, and well, we saw how well Trump did stuff like that.

 

[DigDog]

The staff at Intel may yet be willing to change.

If you are enthusiastic about your work, but you see teh company going in the wrong direction, you'll soon become demoralized and start thinking "that's just the way things go around here" and you'll stop trying to fight for what you think is right.
But something as simple as the CEO saying "we want to change" may be enough motivation for everyone at the same time to have some real effect.

Or, Intel could just let iself die. I'm pretty sure the guys from the financial department have explained how that is not good.

 

[CHADBOGA]

Its not a culture thing its a price issue.
If they had kept prices streamline and not tried to stuff there wallets, they wouldn't have been in such a slump.

Also, they just don't care about listening to us.

Before they had a monopoly on the market, but with Ryzen, people are just fed up and now have a different direction to goto.
This is what broke the camel's back.
If they had these prices at Ryzen Launch, people wouldnt of been in mass Exodus to Camp AMD.

I was an avid pro Intel supporter, even had high branch internal contacts and had to sign NDA's with them, they would hand me cherry picked ES samples to post over at XS and here. I even got my samples before Anandtech did especially with the entire Nehalem arch.

But Intel is definitely not the same company i remember working with.
Anyhow, i really would like to see Intel clean up house and make stuff great again, but it sounds like a Trump campaign, only Make America Great is replaced by Make Intel Great, and well, we saw how well Trump did stuff like that.

Yes, he won the election everyone said he had no hope of winning.

 

[moinmoin]

Its not a culture thing its a price issue.
If they had kept prices streamline and not tried to stuff there wallets, they wouldn't have been in such a slump.

I'm confused, what slump? Because financially (stuffing their wallets) they are successful to a point where stock markets still pretend everything's fine with Intel.

 

[dmens]

The staff at Intel may yet be willing to change.

If you are enthusiastic about your work, but you see teh company going in the wrong direction, you'll soon become demoralized and start thinking "that's just the way things go around here" and you'll stop trying to fight for what you think is right.
But something as simple as the CEO saying "we want to change" may be enough motivation for everyone at the same time to have some real effect.

Or, Intel could just let iself die. I'm pretty sure the guys from the financial department have explained how that is not good.

Too late, the smart people gave up and left for greener pastures. I am sure management is well aware of the brain drain but they don't care, they view it as cost reduction.

 

[AnandThenMan]

The problem Intel has is they are used to having no competition. It's the nature of the beast companies get complacent and milk the market for all it is worth.

 

[VirtualLarry]

The problem Intel has is they are used to having no competition. It's the nature of the beast companies get complacent and milk the market for all it is worth.

If they had kept giving us, the customers, the same die size (roughly) for each new generation, only packed with more and more transistors, giving us more and more cores, and better, higher-performing cores (like AMD has with successive Zen generations), then perhaps Intel might have been in a more competitive position, irregardless of their 10nm "woes".

From what I understand, they kept their cores mostly the same, and the dies that they sold as part of the consumer CPUs, kept getting smaller and smaller, as far as silicon area, with the SAME number of cores (4C/8T), and then Intel simply just kept upping their profit per CPU sold.

They stopped giving good value for their CPUs, and went into complacent, profiteering mode. After all, who needs more than "four cores" (640KB, anyone)?

 

[maddie]

The problem Intel has is they are used to having no competition. It's the nature of the beast companies get complacent and milk the market for all it is worth.

Well, to be honest, companies are artificial constructs to turn "stuff" into money for the investors. We, and I, tend to consider them as creatures with an independent existence and feel emotions when they fail/die. I suppose the failure is us, in feeling that way at all.

 

[beginner99]

The problem Intel has is they are used to having no competition. It's the nature of the beast companies get complacent and milk the market for all it is worth.

This sounds bad but in reality it's actually a good thing. If intel had not milked customers with quad cores for years and actually delivered more cores also on server, AMD would have a lot more trouble gaining market share. It's not like AMD is clearly ahead even performance/$ wise. But people were just fed up with intel.

If intel had 10nm and core count increased on schedule, AMD might haven gone bankrupt before they could catch up. So if the leading company gets complacent, it's a good thing for consumers in the long run. If they keep delvering for over a decade, then any competiont would be dead and a full monopoly would be terrible.

 

[coercitiv]

Intel going complacent is definitely not a good thing for consumers.

These "different timeline" scenarios keep popping up in the forums to illustrate isolated effects (what if <good thing had happened> for Intel?) , but they fail to mention one important aspect: in order for the event opposite to real history to happen, an important piece of the historical puzzle needs to change. Cause and effect, not just effect. If one assumes Intel changes course and develops "immunity" to competition in 2017, then something different needs to happen prior to 2017. Intel is run by a collective mind, hence reacts to more than just isolated personal intervention. Something important needs to go down differently in the history between Intel and AMD in order for this to happen.

It is very likely that this different timeline in which Intel is kept away from complacency may actually be far more lucrative for consumers than what we get to experience.

 

[maddie]

The old saying is always relevant.

The cure for high prices is high prices