Question Intel's future after Pat Gelsinger

[Thibsie]

The decision to juice those to the max must have been made at the top via demand to get highest freq possible, somebody like Gelsinger must have known what the risks are.

Probably, except if the middle management is, again, doing the bad job, here.
In any way, he should have taken responsability for it in the name of the company, which he didn't.

 

[Win2012R2]

Probably, except if the middle management is, again, doing the bad job, here.

They allegedly want to fire 20%, just how many middle managers can possibly be in a company in relative terms?

Unlikely to be 20% that's for sure

 

[Thibsie]

They allegedly want to fire 20%, just how many middle managers can possibly be in a company in relative terms?

Unlikely to be 20% that's for sure

That's Intel we're talking about, who knows....

 

[inquiss]

Like at this point it should be clear to all of us that fabs are only one part of the entire equation. The progression from pentium to core, or bulldozer to zen, both exemplify how its design that really drives the overall performance. Design does have to consider the nodes available yes, but ultimately from a physics standpoint, there's only so many ways to get work done and speed is something you measure. You don't assume the top value of a gauge is the actual performance for most things in life.

Intel has had a massive middle management problem for a while now, and it continues to show. The fact that 13/14th gen chips had degradation issues, and the delay in getting angstrom chips out just goes to show that PG didn't fix the core problems. I truly believe intel has good engineers, good fabs, good designers, but something else is preventing everything from mixing coherently.

I really hope the company gets it together because a world where intel chips are way slower and hotter than the competition is very sad to me. If anything we're quickly approaching an age where intel GPUs are the real gangbuster and most people will be paring them with AMD CPUs, a complete population inversion of just 15 years go. And yet another symptom of total mismanagement.

Intel consumer GPUs are going nowhere. Curious what made you think they are?

 

[dr1337]

Intel consumer GPUs are going nowhere. Curious what made you think they are?

In terms of improvement per year, its the best intel has executed in a long time. Sure, everything is in question now with the new CEO, but if intel could make as much progress in CPU arch as they did in GPU arch, core ultra might have been a consumer hit.

My point is the Xe teams were able to execute despite graphics NOT being a focus for intel. The only reason they aren't executing as well in CPU is not due to problems in node but entirely due to the CPU teams being mismanaged.

 

[johnsonwax]

They stopped being paranoid

They were comfortably ahead of global foundries (& bulldozer)

But starting jaguar, bobcat (& zen) AMD pulled ahead with TSMC partnership

Right, but that TSMC partnership is not independent of what I said. ARM + foundries blew holes in both of Intel's moats - proprietary x86 tied to the best process - which reinforced each other. Once enough dollars flowed into foundries like TSMC via mobile, they caught up quickly on process, and because AMD could also sell x86 they could chase competitive processes. And because ARM was licensable, a thousand flowers bloomed in that space for more narrowly tailored silicon than what a monolithic architecture could cover. AMD took advantage of the funding opportunity to TSMC that those competitors brought, but that also allowed Apple to later dump Intel, and then to some degree Microsoft.

It wasn't just paranoia they needed, they had the entirely wrong business model for this new environment. And that goes back at least as far as Intel failing to get Apple's mobile business years earlier, though nobody could have seen that at the time. They weren't in a position to let x86 go, or to open up their foundry business, and they rode that horse too long, because it was the only horse they knew how to ride. Changing business model is so hard that it almost never happens, until the business bottoms out and they have no choice. Now they have no moats, no meaningful advantages, and the wrong business model. And the board is giving each CEO more rope, and yet not enough to actually turn the ship.

 

[johnsonwax]

They allegedly want to fire 20%, just how many middle managers can possibly be in a company in relative terms?

Unlikely to be 20% that's for sure

20% isn't hard to hit, tbh. Particularly when you are acquiring and throwing off so many new businesses as Intel did. You build up these layers to mange the new complexity but you don't fully remove them when you cast that complexity off.

 

[johnsonwax]

In terms of improvement per year, its the best intel has executed in a long time. Sure, everything is in question now with the new CEO, but if intel could make as much progress in CPU arch as they did in GPU arch, core ultra might have been a consumer hit.

My point is the Xe teams were able to execute despite graphics NOT being a focus for intel. The only reason they aren't executing as well in CPU is not due to problems in node but entirely due to the CPU teams being mismanaged.

The reason why both AMD and Intel are losing to Apple on performance potential is not because of management or engineering, but lack of control. Apple is able to shed a shit ton of technical debt because they can force changes on the software/OS side (ensure backward compatibility through software freeing up changes in hardware), on the developers (we're killing 32 bit, get over it), and so on. FFS, they've changed ISA 3 times. AMD/Intel generally lack that ability, and you can't reorganize into solving that problem. Apple on the compatibility vs performance trade off sides with performance, while Intel/AMD side with compatibility, and that's not an easy thing for them to change.

GPU has WAY less baggage.

 

[Joe NYC]

Probably, except if the middle management is, again, doing the bad job, here.
In any way, he should have taken responsability for it in the name of the company, which he didn't.

Since you can't blame the workers and can't blame the current CEO + upper management, "middle management" remains a convenient scape goat.

 

[Kepler_L2]

Lip-Bu Tan: Our Path Forward

www.intc.com

 

[fastandfurious6]

good decisions except RTO

 

[Thunder 57]

good decisions except RTO

I agree. If three days a week being enforced is "uneven", enfore it first before you go to four days. You are already cutting a massive number of jobs and now you want them in the office more? Ever hear of burnout?

 

[gdansk]

good decisions except RTO

It's part of the cost cutting. It'll get rid of a few. Hope they aren't the best.

 

[Tuna-Fish]

Many teams are eight or more layers deep

6+ layers of middle management?

 

[Hitman928]

6+ layers of middle management?

Intel is notorious for having managers for managers managing managers. Worse yet, I’ve heard from multiple former employees that most of that middle management is very light on technical skills/experience.

 

[fastandfurious6]

Intel is notorious for having managers for managers managing managers. Worse yet, I’ve heard from multiple former employees that most of that middle management is very light on technical skills/experience.

what's the point 😭 where did this even begin from?

what depths of hell conceived this nightmare on earth...

god forgive us 😔

 

[igor_kavinski]

Intel is notorious for having managers for managers managing managers. Worse yet, I’ve heard from multiple former employees that most of that middle management is very light on technical skills/experience.

Sounds exactly like my company coz the HR head is a complete idiot.

 

[igor_kavinski]

It's part of the cost cutting. It'll get rid of a few. Hope they aren't the best.

Hopefully this will weed out at least a hundred powerpoint monkeys that Raja snuck into Intel

 

[igor_kavinski]

Intel uncovers alleged embezzlement involving former employee and supplier

An insider and an outside allegedly colluded to embezzle over $840,000 from Intel.

www.tomshardware.com

The less said the better.

 

[511]

well the structure is changing they said bye bye to many people :rofl:.

 

[igor_kavinski]

well the structure is changing they said bye bye to many people :rofl:.

Hey, I'm available for the position of Director of Innovation. I'll bring back P4 7 GHz. Intel will be back at the top, at least in terms of GHz. Who cares about performance? Even Core i3 has enough performance for 4K resolution gaming!

 

[511]

Hey, I'm available for the position of Director of Innovation. I'll bring back P4 7 GHz. Intel will be back at the top, at least in terms of GHz. Who cares about performance? Even Core i3 has enough performance for 4K resolution gaming!

i will be even better pick than you than i will be bringing back 10Ghz Tejas.

 

[igor_kavinski]

i will be even better pick than you than i will be bringing back 10Ghz Tejas.

Nuh uh. They had a working 7 GHz prototype. Just revive that easy peasy. If you go for 10 GHz, you will be no better than Raja selling snake oil of 100 GAJILLION TOPS!

 

[reaperrr3]

You are already cutting a massive number of jobs and now you want them in the office more? Ever hear of burnout?

I mean, cutting many jobs means offices will be a lot emptier, unless they bring more remaining workforce back on-site...

But I also believe companies like Intel likely have some data on how home office stacks up compared to being on-site in terms of getting work done, or impact on team-work in general.
Sure, there will be workers who work just as efficiently or even more efficiently from home, but there will also be a lot where the opposite is the case.
I think companies wouldn't push for RTO so hard against workers' wishes unless data indicated that too many workers somehow slack off at home in some form, or that teamwork degrades etc.

As for burnout, it's unlikely that 1 day more of commuting and meeting colleagues in person will burn someone out, unless they've already been on the brink anyway.
And as harsh as it sounds, Intel needs the entire remaining workforce to really roll up their sleeves now, or the entire ship will go down sooner rather than later and none of it will matter anymore.

Besides, workers' working conditions at Intel today are probably still a lot more convenient compared to AMD 2012, for example.

 

[Doug S]

Worse yet, I’ve heard from multiple former employees that most of that middle management is very light on technical skills/experience.

That's a problem everywhere. Even if you make a goal to have managers with strong technical skills and experience the people who fit that bill often don't WANT to be spending most of their time doing management tasks.

I could tell a long story about how I had my first management job at 30, hated it and quit after a few years to become a full time consultant as escape the management track, only to increasingly find that in the architectural type role I'd slowly moved into over the years that even as a consultant I ended up dealing with a lot of the stuff I hated about being a manager and retired at 51 because it just wasn't any fun anymore.

The companies that have a lot of strong technical people in their ranks are generally those that have grown a lot in the past and they're locked in place by unvested stock options. All the unvested stock options at Intel are underwater, so there was nothing keeping them in place. Not sure if Intel had open offers for long time (usually 20 years) employees to take early retirement, but if so that's probably where most of the technical people who didn't like management went. The ones who stayed are the ones who like being a manager, and even if they used to be technical they happily forgot all that when they found they love spreadsheets and Powerpoint more!