Former Intel CEO on What Went Wrong, What's Next + Lovable CEO on the Real Promise of Vibe Coding

[Daniel Nenni]

Former Intel CEO Pat Gelsinger joins Jason!
- What Went Wrong at Intel
- Why a Taiwan Blockade Would Cripple the US Economy
- Lovable's Anton Osika: One Million New Apps a Week
- How Lovable is Bringing Down Builder Costs

Probably one of the greatest American companies ever. And then absolutely went off the rails and got absolutely demolished by Nvidia, TSMC, and I guess Apple to a certain extent.

 

[hist78]

View attachment 4885

Former Intel CEO Pat Gelsinger joins Jason!
- What Went Wrong at Intel
- Why a Taiwan Blockade Would Cripple the US Economy
- Lovable's Anton Osika: One Million New Apps a Week
- How Lovable is Bringing Down Builder Costs

Probably one of the greatest American companies ever. And then absolutely went off the rails and got absolutely demolished by Nvidia, TSMC, and I guess Apple to a certain extent.

Why does Pat Gelsinger keep saying he came back to Intel as CEO in 2001 instead of 2021? Is he okay?

 

[Daniel Nenni]

Why does Pat Gelsinger keep saying he came back to Intel as CEO in 2001 instead of 2021? Is he okay?

Has he ever been a great communicator? It seems his truth may not match ours.

 

[MKWVentures]

What went wrong is simple:
Pat said there were foundry customers and there would be more. after 4+ years there were no customers at all
Pat said 20A and 18A were doing well and Intel products and external customers would jump to use them. They were not doing well
Pat said Intel needed to build 6 new fabs to keep up with internal and external demand. Only one was needed. and it was delayed over a year and it is still not fully loaded as of today

the result of the above was literally a dangerous balance sheet, hence the stock price below book.

All of the above was caused by Pat. Without Pat, they might have lost market share as well but they wouldn't have the finanicial disaster that LBT fixed.

Side note: Pat also continued and increased the arrogance of Intel being the greatest company ever and people are lucky to buy from Intel. For some reason customers didnt like this (LBT has great stories along this line.... so do I).

Intel still needs to cost effectively execute on the plan and it is not clear what will happen. But LBT was needed to correct the issues.

FYI: I am stunned to see how much change in operations, culture, management, finance decisions LBT is has made in 18 months. Amazing.... I have never seen changes like this (Maybe TI many years ago)

 

[blueone]

Why does Pat Gelsinger keep saying he came back to Intel as CEO in 2001 instead of 2021? Is he okay?

To clarify, Gelsinger's gaff is not that he became CEO in 2001, it is that he became Intel's first corporate CTO in 2001. He misspoke about his title. IMO, he was a disaster as a CTO too. (I was there at the time.)

 

[MKWVentures]

To clarify, Gelsinger's gaff is not that he became CEO in 2001, it is that he became Intel's first corporate CTO in 2001. He misspoke about his title. IMO, he was a disaster as a CTO too. (I was there at the time.)

100% agree. I do not recall PG being terribly successful or very well respected in any of his intel rolls. His positive is relentless optimism.... which I saw in his GM/VP roles. It would be wrong to say he "led" technology as CTO. I would say he was a cheerleader and spokesman for other peoples work.

 

[Daniel Nenni]

I honestly liked Pat at the beginning but he did not have control over his mouth and he still had the Intel bravado that grew stale years ago. He insulted TSMC and TSMC customers. Now he is badmouthing Taiwan while holding manufacturing assets in Israel who is actually at war.

I remember when GlobalFoundries came out and did the same thing. Taiwan is at risk! The Semiconductor sky is falling! Yet here we are with TSMC in the most dominant position ever in the history of semiconductors. And where is Globalfoundries? 99% of the leading edge semiconductors come from TSMC and will continue to do so at 2nm and maybe even 1.4nm.

The same thing could be said about memory. South Korea controls memory and has a crazy neighbor to the North that wants to annihilate them.

China is not stupid. They are not semiconductor independent so any actions against Taiwan will hurt the entire world including China. In fact, I think Taiwan is safer now than ever before. The Silicon Shield is the strongest it has ever been. Semiconductors are front page news. Everyone in power knows how important they are to modern life and will defend the supply chain. Otherwise we will enter a new version of the dark ages that the world population will not like at all, absolutely.

 

[blueone]

100% agree. I do not recall PG being terribly successful or very well respected in any of his intel rolls. His positive is relentless optimism.... which I saw in his GM/VP roles.

He was the chief architect for the 80486 CPU, which directly or indirectly won him various corporate and industry accolades and awards, including being named an IEEE Fellow and a member of the National Academy of Engineering. I wasn't in Intel at the time, but later I heard a variety of opinions about his technical contributions for the 80486, good and bad.

It would be wrong to say he "led" technology as CTO. I would say he was a cheerleader and spokesman for other peoples work.

The truth is, in many large companies the title CTO really stands for Chief Talking Officer. The real decisions are made by management executives. CTO is often an out-bound role rather than an internal leadership role. I know; I once thought I accepted a job as Chief Technical Officer, only to find out I was really just supposed to talk to customers and do industry presentations. The interviews were all about what I thought should be technology strategies, fooling me. It didn't end well.

 

[Daniel Nenni]

"I think one of the fundamental things is, and as you look at the great technology companies today, they're deeply technical," Gelsinger told "All-In Podcast" cohost Jason Calacanis during an interview taped at the Raise Summit in Paris.

Gelsinger said when he returned to Intel in 2021, he "was the first technical leader in essentially 15 years associated with it." The former Intel chief said his predecessors had the wrong background.

"When you're making these hardcore technical decisions that affect billions of dollars, you don't do that through a spreadsheet," Gelsinger said. "That's a lousy investment unless the technology trends make it the right investment."

Paul Otellini, who led Intel from 2005 to 2013, was the first non-engineer to lead the company, coming from the business side. Brian Krzanich, who has a degree in chemistry, replaced Otellini after working his way up from his start at a chip factory in New Mexico, and later spent decades working on manufacturing issues. Bob Swan, the final executive before Gelsinger took over, had spent decades in top financial positions, including as CFO of eBay and HP Enterprise Services.

Gelsinger also took issue with the amount of money Intel spent on dividends and shareholder buybacks before he took over. According to Intel's financial filings, the company sent back roughly $79 billion to shareholders through stock buybacks and dividends from 2015 through 2020.

"What I wouldn't have done for another hundred billion dollars on the balance sheet?" he said.

Former Intel CEO says the chipmaker went off the rails ‘when it started to be run by business people’

Former Intel CEO Pat Gelsinger said for over a decade the chipmaker was run by the wrong type of executives.

www.yahoo.com

 

[Xebec]

The truth is, in many large companies the title CTO really stands for Chief Talking Officer. The real decisions are made by management executives. CTO is often an out-bound role rather than an internal leadership role. I know; I once thought I accepted a job as Chief Technical Officer, only to find out I was really just supposed to talk to customers and do industry presentations. The interviews were all about what I thought should be technology strategies, fooling me. It didn't end well.

+1 on the CTO observation - it is often "intended" as a customer (internal or external) / public relations role instead of actually technical.

"Engineering VPs" (and Directors) are usually the important "technical leadership" execution roles in large companies. (Of course bad technical leaders can still be .. overcome by great technical team members/contributors).

Though, there are exceptions -- John Carmack was CTO at Meta for the Quest/Oculus products, and while he was responsible for hype he also made sure that a ton of latency issues were addressed with VR products in general. John's expertise led to software and hardware hitting much harder than it would have otherwise, while he still did the public appearances expected of the role.

 

[Xebec]

He was the chief architect for the 80486 CPU, which directly or indirectly won him various corporate and industry accolades and awards, including being named an IEEE Fellow and a member of the National Academy of Engineering. I wasn't in Intel at the time, but later I heard a variety of opinions about his technical contributions for the 80486, good and bad.

As a total Intel outsider, he honestly looked to me like a great candidate for the CEO role. He was mentored directly by one or more Founders (I always hold Founders in regard above any other corporate executive). He had the appearance of technical chops that several predecessors lacked. Intel needed energy as well at the top, rather than just "these are our quarterly results, and 10nm has been delayed again" which went on for at least 5 years.

He was also handed a really bad deck of cards by the Board of Directors (who approved his strategy change):
- Intel had already switched to outsourcing strategy, poisoning any future move back to in-sourcing financially. ("TSMC poison pill").
- Intel's execution on process nodes was failing catastrophically
- Intel's execution on server architecture was largely failing
- Intel had already given up several markets that were now competing with them: Mobile, Apple, and "GPU applications"
- Intel's next attempt at a GPU product was extremely poorly executed upon*, meaning a multi-year loss-making product was handed to Pat
- Intel's financials were weakening significantly as AMD was surging
- 3D XPoint was a looming financial disaster handed to Pat by previous CEOs

However - he certainly made a LOT of forced errors as CEO:
- Overspending without having customers lined up (or at least not committed)
- Not doing what was necessary to secure more CHIP Act (v1.0) money for Intel to offset some of this spending
- Handling the Raptor Lake recall situation very poorly (massive reputation loss among smaller scale server providers, and retail customers/OEMs)
- Not acting aggressively enough to fix the server product line (though hard to tell how bad this really was at the beginning)
- The Tower semiconductor acquisition could have helped cross the bridge on some customer commitments, but that also counts as a Pat failure since it was his responsibility to make it happen.
- PR Disasters in a few cases including comments about customers/partners that weren't well received / not polished like they should be from a CEO

.. Anyway, I simply wish he (and Intel) did better. In 20 years if Intel is gone, it'll be interesting to see whether it was him or Otellini that ultimately cost Intel it's future. I'm also not sure 18A would exist at all if Pat hadn't become CEO - is that good or bad?.

P.S. My perspective on Otellini's time as CEO: Paul first gave up on ARM internally (StrongARM sold in 2006), and then surrended the mobile market to ARM. Paul also failed to execute on building GPUs and early AI efforts. Each of these markets are proving larger than Intel's core business today. IMO, Paul also failed to kill off AMD when there was an opportunity.

*Raja Koduri should have done a lot better here - I feel strongly (based on interviews) that ARC Alchemist had basic architectural mistakes that someone as experienced as Raja should have caught early.
-

 

[tim_b]

As a total Intel outsider, he honestly looked to me like a great candidate for the CEO role. He was mentored directly by one or more Founders (I always hold Founders in regard above any other corporate executive). He had the appearance of technical chops that several predecessors lacked. Intel needed energy as well at the top, rather than just "these are our quarterly results, and 10nm has been delayed again" which went on for at least 5 years.

He was also handed a really bad deck of cards by the Board of Directors (who approved his strategy change):
- Intel had already switched to outsourcing strategy, poisoning any future move back to in-sourcing financially. ("TSMC poison pill").
- Intel's execution on process nodes was failing catastrophically
- Intel's execution on server architecture was largely failing
- Intel had already given up several markets that were now competing with them: Mobile, Apple, and "GPU applications"
- Intel's next attempt at a GPU product was extremely poorly executed upon*, meaning a multi-year loss-making product was handed to Pat
- Intel's financials were weakening significantly as AMD was surging
- 3D XPoint was a looming financial disaster handed to Pat by previous CEOs

However - he certainly made a LOT of forced errors as CEO:
- Overspending without having customers lined up (or at least not committed)
- Not doing what was necessary to secure more CHIP Act (v1.0) money for Intel to offset some of this spending
- Handling the Raptor Lake recall situation very poorly (massive reputation loss among smaller scale server providers, and retail customers/OEMs)
- Not acting aggressively enough to fix the server product line (though hard to tell how bad this really was at the beginning)
- The Tower semiconductor acquisition could have helped cross the bridge on some customer commitments, but that also counts as a Pat failure since it was his responsibility to make it happen.
- PR Disasters in a few cases including comments about customers/partners that weren't well received / not polished like they should be from a CEO

.. Anyway, I simply wish he (and Intel) did better. In 20 years if Intel is gone, it'll be interesting to see whether it was him or Otellini that ultimately cost Intel it's future. I'm also not sure 18A would exist at all if Pat hadn't become CEO - is that good or bad?.

P.S. My perspective on Otellini's time as CEO: Paul first gave up on ARM internally (StrongARM sold in 2006), and then surrended the mobile market to ARM. Paul also failed to execute on building GPUs and early AI efforts. Each of these markets are proving larger than Intel's core business today. IMO, Paul also failed to kill off AMD when there was an opportunity.

*Raja Koduri should have done a lot better here - I feel strongly (based on interviews) that ARC Alchemist had basic architectural mistakes that someone as experienced as Raja should have caught early.
-

My impression is that he has the capability to manage a business the size and complexity of VMware or EMI without creating a disastrous outcome. However, anything significantly larger seems to be beyond his comfort zone. It's like someone who can confidently drive a golf cart—handing them the keys to an 18-wheeler is a completely different challenge, and the results are unlikely to be good.

 

[blueone]

My impression is that he has the capability to manage a business the size and complexity of VMware or EMI without creating a disastrous outcome. However, anything significantly larger seems to be beyond his comfort zone. It's like someone who can confidently drive a golf cart—handing them the keys to an 18-wheeler is a completely different challenge, and the results are unlikely to be good.

I think it's more complexity than size. The problem an IDM like Intel has is that finding the one CEO that can have the decision-making capabilities for product design and for semiconductor manufacturing is pretty much a null set, in my experience. If ever there was a case for co-CEOs, it's Intel. Grove, Noyce, and Moore (and Barrett) wouldn't do much better in my opinion either. Being deeply technical in materials science or chemical engineering is far different than being deeply technical in computer engineering. I've never heard of anyone or met anyone who could be capable at both. (Are you listening, Elon? )

 

[tim_b]

I think it's more complexity than size. The problem an IDM like Intel has is that finding the one CEO that can have the decision-making capabilities for product design and for semiconductor manufacturing is pretty much a null set, in my experience. If ever there was a case for co-CEOs, it's Intel. Grove, Noyce, and Moore (and Barrett) wouldn't do much better in my opinion either. Being deeply technical in materials science or chemical engineering is far different than being deeply technical in computer engineering. I've never heard of anyone or met anyone who could be capable at both. (At you listening, Elon? )

If someone lacks expertise in a particular area, they can always delegate to people who have deeper knowledge. The challenge isn't necessarily that a CEO needs to be world-class at both semiconductor manufacturing and product design. The real test is knowing where your own limits are, identifying the right experts, and making sound decisions based on their input.

In Pat's case, I think the problem was less about a lack of technical capability and more about judgment. He appeared to put a great deal of confidence in the wrong people and the wrong strategies. Delegation only works when you choose the right people to delegate to and maintain enough skepticism to challenge their assumptions. If your advisors are weak—or if you trust them too much without adequate scrutiny—you can still end up making poor decisions regardless of your own background.

 

[blueone]

If someone lacks expertise in a particular area, they can always delegate to people who have deeper knowledge. The challenge isn't necessarily that a CEO needs to be world-class at both semiconductor manufacturing and product design. The real test is knowing where your own limits are, identifying the right experts, and making sound decisions based on their input.

In theory, yes. In reality, I can't think of a case where the theory played out well.

In Pat's case, I think the problem was less about a lack of technical capability and more about judgment. He appeared to put a great deal of confidence in the wrong people and the wrong strategies. Delegation only works when you choose the right people to delegate to and maintain enough skepticism to challenge their assumptions. If your advisors are weak—or if you trust them too much without adequate scrutiny—you can still end up making poor decisions regardless of your own background.

Pat is about his personal agenda. Pat was never one to have advisors. He likes people who execute his agenda. He reminds me of a certain US president.

 

[blueone]

"I think one of the fundamental things is, and as you look at the great technology companies today, they're deeply technical," Gelsinger told "All-In Podcast" cohost Jason Calacanis during an interview taped at the Raise Summit in Paris.

I agree with this statement.

Gelsinger said when he returned to Intel in 2021, he "was the first technical leader in essentially 15 years associated with it." The former Intel chief said his predecessors had the wrong background.

The best CEO I ever worked under in a technology company was someone with a finance background who hired highly technical people with a vision. Having a PhD in materials science does not necessarily make you technical in various chip and system design fields. Itanium was born under Barrett.

There's a phenomenon I have observed over the decades, which I gave the name: "Smartest Person in the Room Syndrome". The subject here is almost universally male. In one or perhaps a few corporate "rooms" these people are indeed the smartest person in the room. Sometimes a recognized industry expert. After a while they get to be more senior and they're in more rooms, and some develop the delusion that because they are the smartest person in one of a few rooms, they are the smartest person in whatever room they happen to be in. In the new rooms, where their perceived expertise is mostly based on their seniority, they really aren't deep experts. They think they can understand technical complexity with a few days of superficial studying, so they really don't know what they're talking about, and make fools of themselves (and make poor decisions). The irony is that people afflicted with the syndrome are usually scathing and intolerant with people posing as experts in their "home rooms".

I've always had the most respect for senior people who had deep expertise, but know what they don't know. I think that quality is relatively rare.

"What I wouldn't have done for another hundred billion dollars on the balance sheet?" he said.

Former Intel CEO says the chipmaker went off the rails ‘when it started to be run by business people’

Former Intel CEO Pat Gelsinger said for over a decade the chipmaker was run by the wrong type of executives.

www.yahoo.com

I'm sure he would have just poorly invested another $100B.

 

[swka]

I agree with this statement.

The best CEO I ever worked under in a technology company was someone with a finance background who hired highly technical people with a vision. Having a PhD in materials science does not necessarily make you technical in various chip and system design fields. Itanium was born under Barrett.

There's a phenomenon I have observed over the decades, which I gave the name: "Smartest Person in the Room Syndrome". The subject here is almost universally male. In one or perhaps a few corporate "rooms" these people are indeed the smartest person in the room. Sometimes a recognized industry expert. After a while they get to be more senior and they're in more rooms, and some develop the delusion that because they are the smartest person in one of a few rooms, they are the smartest person in whatever room they happen to be in. In the new rooms, where their perceived expertise is mostly based on their seniority, they really aren't deep experts. They think they can understand technical complexity with a few days of superficial studying, so they really don't know what they're talking about, and make fools of themselves (and make poor decisions). The irony is that people afflicted with the syndrome are usually scathing and intolerant with people posing as experts in their "home rooms".

I've always had the most respect for senior people who had deep expertise, but know what they don't know. I think that quality is relatively rare.

I'm sure he would have just poorly invested another $100B.

did not know Pat is one of those "smartest in the room" one. Certainly does not look like that from outside looking in.

From my experience, those people typically have a high sense of insecurity, and has to constantly overshadow others to remind everyone else their status

 

[user nl]

Nothing new in that podcast, no reflection on his own leadership at INTEL. I learned nothing new, just confirming that PatG likes to be a cheerleader, just a bubble of talk. Not sure he is a wise leader.


The guy who interviewed him when he was 18 for a job at INTEL wrote in his report on PG:

In 1979, he was interviewed at the technical institute by a manager from Intel. Unlike most of the other students there, Mr. Gelsinger had heard of the company. He breezed through questions related to his studies and predicted he could earn bachelor’s, master’s and Ph.D. degrees while holding down a full-time job, said Ronald Smith, the former Intel executive who conducted the interview.

“He is very smart, very ambitious and arrogant,” Mr. Smith said he wrote in a summary of the conversation. “He’ll fit right in.”
..............................................................................................

Patrick Gelsinger was 18 years old and four months into an entry-level job at Intel when he heard a pivotal sermon at a Silicon Valley church in February 1980. There, a minister quoted Jesus from the Book of Revelation.

“I know your deeds, that you are neither cold nor hot. I wish you were either one or the other!” the minister said. “So, because you are lukewarm — neither hot nor cold — I am about to spit you out of my mouth.”

The words jolted Mr. Gelsinger, reshaping his philosophy. He realized he had been a lukewarm believer, one who practiced his faith just once a week. He vowed never to be neither hot nor cold again.

https://www.nytimes.com/2022/02/17/technology/intel-ceo-patrick-gelsinger.html