Support NeoGAF

[winjer]

The AI War of The Words: Intel CEO Says NVIDIA's Success Came From "Pure Luck", NVIDIA Says It Came From Vision & Execution Which Intel Lacked

Intel & many other tech companies have amassed against NVIDIA in a "War of Words" claiming that the green team's AI success came from "Luck".

wccftech.com

Jensen worked super hard at owning throughput computing, primarily for graphics initially, and then got extraordinarily lucky. They didn't even want to support their first AI project. Pat Gelsinger (Intel CEO) via Manufacturing@MIT

The interesting part here is that apparently, Intel's CEO believes that if the company had continued its Larabee project, which was ongoing until Gelsinger left the company, and then later returned 11 years ago.

For those unaware of Intel Larabee, it was a project initiated by Intel to develop a highly parallel, many-core architecture. It was initially intended for the consumer segment, however, Intel shifted its focus and repositioned Larabee for high-performance computing (HPC) and parallel processing tasks. Team Blue didn't see the AI hype a decade ago, but the Larabee project could've positioned them better in the markets.

However, the twist of the plot here is that NVIDIA's Vice President of Applied Deep Learning Research, Bryan Catanzaro, and a former Intel employee have revealed that Intel's Larabee project and the whole idea of AI-accelerated computing lacked focus and execution from the firm, which is why it isn't in the position the company could've been. Bryan reveals that despite having 10x larger revenues than NVIDIA, Intel at that time was unsure of what they were doing with Larabee, which ultimately had put them behind Team Green fast-forwarding a decade. With ongoing GPGPU and compute efforts, both hardware and software, NVIDIA conquered the market with no competition in sight.

Seems like someone is very envious of Nvidia's success in the AI market. And that NVidia today is valued a lot higher than Intel.

 

[BlackTron]

 

[poppabk]

Nvidia gave been lucky that GPU processing power found other uses in crypto and AI, I don't think that was the plan, and if GPU's were still just for graphics they would be a fraction of the size they are today.

 

[Dirk Benedict]

Nvidia is evil. Fuck this company. Even if I buy their shit.

 

[E-Cat]

Intel’s current position came from extraordinary stupidity.

 

[ahtlas7]

I thought it was because of

 

[Rentahamster]

Intel probably should have tried innovating instead of resting on their market leading laurels for 10 years.

 

[LiquidMetal14]

This says a lot. I like Bryan as well and has a lot more equity with me than Pat. Pat is sufferingfrom the same thing Uncle Phillis Spender from Microsoft is suffering from and it's not success.

 

[winjer]

This says a lot. I like Bryan as well and has a lot more equity with me than Pat. Pat is sufferingfrom the same thing Uncle Phillis Spender from Microsoft is suffering from and it's not success.

That as during Otellini, a guy with no technical know how, that only cared about marketing.
At this point Intel was still a tech giant, but he took the first steps for Intel to lose it's competitiveness in tech.
And then came the 1-2 punch, and with Krzanich, investments into R&D were cut even lower.
In a decade, these two squandered the big advantage Intel had over everyone else.

Gelsinger doesn't have aa easy task ahead of him.

 

[LiquidMetal14]

That as during Otellini, a guy with no technical know how, that only cared about marketing.
At this point Intel was still a tech giant, but he took the first steps for Intel to lose it's competitiveness in tech.
And then came the 1-2 punch, and with Krzanich, investments into R&D were cut even lower.
In a decade, these two squandered the big advantage Intel had over everyone else.

Gelsinger doesn't have aa easy task ahead of him.

Agree.

I think Pat and his staff need to put their heads down and work. Their blunders have gone far and are of their own doing. They are the lucky ones to have the cash to spend while poorly handling the consumer end alongside lackluster performance vs the competition.

I had investment in Intel for some years a while ago and continued to be a supporter of their desktop parts for over 10+ years. I hold no bias much anymore hence me being primarily AMD for the last 2-3 years.

 

[LordOfChaos]

So many publications apparently turned tabloids trying to spin this for clicks.

What Pat said was Jen-Hsun himself told him that he felt like they lucked out with AI. They work together all the time particularly with foundry services coming up and products going back and forth you know.

It's in here. This is a vastly different context than the arrogance they tried to spin it as. Same with his comments about ARM being relatively small on Windows right now, what he said was that they keep their eyes on everything.

 

[LiquidMetal14]

So many publications apparently turned tabloids trying to spin this for clicks.

What Pat said was Jen-Hsun himself told him that he felt like they lucked out with AI. They work together all the time particularly with foundry services coming up and products going back and forth you know.

It's in here. This is a vastly different context than the arrogance they tried to spin it as. Same with his comments about ARM being relatively small on Windows right now, what he said was that they keep their eyes on everything.

Ugh, thanks for sharing. Hate it when it gets spun wrong.

 

[Yoda]

And what was the cause of his company's current position? Being lazy? Hiring too many MBAs?

 

[Unknown Soldier]

LMAO @ Intel CEO

If the whole company has the same attitude and culture as this CEO, it's no wonder Intel is a dying company

 

[dave_d]

Intel’s current position came from extraordinary stupidity.

Kind of like Sears? I mean I heard someone point out they had everything to be where Amazon is now and have done it about 20-25 years ago. (A retail market and tons of distribution centers. They just needed a web site and some trucks) They're basically out of business at this point. (amazing since when I was a kid they were huge)

 

[TransTrender]

Intel really is fucked in the head

 

[mclaren777]

Project Offset randomly popped into my head yesterday. I used to think it looked incredible, but a lot has changed in the past 14 years.

*context in case you need it*

 

[winjer]

So many publications apparently turned tabloids trying to spin this for clicks.

What Pat said was Jen-Hsun himself told him that he felt like they lucked out with AI. They work together all the time particularly with foundry services coming up and products going back and forth you know.

It's in here. This is a vastly different context than the arrogance they tried to spin it as. Same with his comments about ARM being relatively small on Windows right now, what he said was that they keep their eyes on everything.

You are wrong, Pat did say that NVidia was lucky. This is the presentation at MIT, where he said those words.

Here is a transcript of what he said:

I wanted to be the CEO uh of Intel but when I was pushed out of the company 13 years ago.
They killed the project that would have changed the shape of AI. I had a project underway it was called Larabee at the time to do high throughput, computing in the x86 architecture
They killed it after I left. Nvidia had unmerited potential in the space of AI Hardware, essentially throughput Computing versus scaler Computing.
Nobody put pressure on him. And Jensen by the way, he worked super hard at owning throughput Computing primarily for graphics initially and then he got extraordinarily
lucky. I mean they didn't even want to support their first AI project right and it's just big flop.

 

[LordOfChaos]

You are wrong, Pat did say that NVidia was lucky. This is the presentation at MIT, where he said those words.

Here is a transcript of what he said:

And here's what he said in the one I shared

Interviewer: "Interesting, and obviously it worked really well for gaming, and that's Nvidia's specialty. Is that how Nvidia just ended up running away with the game here, that they built this GPU for gaming and they lucked into it being good for AI?"

Pat: "Yeah, it very much is that way, and Jensen and I have known each other for 35 years. We have this general-purpose workload with the CPU, and we always are adding more capabilities. But over there it was always just 'go really fast' for graphics. And then you got really lucky that the AI workload looked a lot like the graphics workload. So, as I joke with Jensen, I said, 'You just were true to the mission of throughput computing and graphics, and then you got lucky on AI,' and he said, 'No, no, no, Pat. I got really lucky on AI.'"

So this is literally what Jen-Hsun told him according to him

 

[winjer]

And here's what he said in the one I shared

Interviewer: "Interesting, and obviously it worked really well for gaming, and that's Nvidia's specialty. Is that how Nvidia just ended up running away with the game here, that they built this GPU for gaming and they lucked into it being good for AI?"

Pat: "Yeah, it very much is that way, and Jensen and I have known each other for 35 years. We have this general-purpose workload with the CPU, and we always are adding more capabilities. But over there it was always just 'go really fast' for graphics. And then you got really lucky that the AI workload looked a lot like the graphics workload. So, as I joke with Jensen, I said, 'You just were true to the mission of throughput computing and graphics, and then you got lucky on AI,' and he said, 'No, no, no, Pat. I got really lucky on AI.'"

So this is literally what Jen-Hsun told him according to him

I'm not contesting that Jensen told him that.
I'm contesting that you said this was all a misinterpretation by some journalists.
In that presentation at MIT, Pat clearly states that NVidia's successes in AI are unmerited and of extreme luck.
And then we have Bryan Catanzaro, from NVidia, directly contesting Pat's statements.
This is not just tabloid kerfuffle for the sake of it.