It is a hypothetical scenario. If Sam Altman returns Open AI as CEO and its chief scientist Ilya leaves the company, Intel should try to recruit him. The only moat in AI is compute. As Intel is vertically integrated from factories to software, it already has the moat. As Ilya wants to build AGI that benefits all of humanity, which is in alignment with PG, he can become the chief AI scientist at Intel working towards AGI. Intel should also start developing software/service portfolios/business lines.
14 replies (most recent on top)
Is he good at PPT? Don't think so. Not a good fit.
Why pay for another over rated highly paid data scientist to hop on the band wagon with a short run... It'll be funny when all these investments for AI come in and the reality sinks in later about the costs for maintain such large models outweigh the benefits of classic approaches. Similar to the cloud repatriation phenomena happening currently. A bit of buyers remorse.
Why is cost a big concern?
Because everyone wants Nvidia! Duh!
They could have lower costs by buying Intel or AMD but people don’t seem to be using them for much.
I wonder if there’s a reason for that?
It’s almost as if the value nvidia provides is worth the cost? Nah, that couldn’t be it. Nvidia is worth over 1.2 trillion because people are too d-mb to see the value Intel brings to the table.
@1uyo+1pFksPqR if what you say is true, why do so many companies start to do AI chips? I wonder how many commentators here really work on AI models? Do you really know their pain points? The dominant pain point for many researchers and companies is cost. Software support is also a very important factor. Intel has the resources to address those two points. That is the edge.
If people really want those Intel technologies, Intel should have some pricing power there, but it does not.
Instead, people are bidding up Nvidia solutions since there is no viable alternative.
Sorry but you can’t just unseat an incumbent that has built their ecosystem over 15 years. There is too much inertia, much like how wintel won in the past.
@1isl+1pFksPqR My point is that as all those service companies want to build AI chips, Intel already has those chips and software (Xeon, Gaudi, Arc GPU, Falconshore, Open API). For example, Intel can develop AI services that leverage on X86 ecosystems. For example, Meteorlake can use local NPU to do simple AI tasks but it can also access services from the cloud. In this case, the cloud can be Intel cloud or Intel powered cloud with Intel's own AI models. As Intel has its factories, the cost base for those chips is as competitive as other companies. It also have sufficient supply for training models. That is the edge.
Can you elaborate on what strategic assets Intel has that others do not?
You say these words as if they have meaning but it’s actually gibberish unless you give concrete examples.
Process lead - nope.
Hardware lead - nope
Software lead - truly laughable
Ecosystem? What ecosystem?
Look, Intel is good a making x86 processors and at one time in the PAST had a virtuous cycle that allowed them to create a process technology lead through sheer economies of scale.
That lead was vaporized once smartphones took over the leading edge volumes. The ‘open’ TSMC ecosystem ran over Intel primarily due to smartphones. That’s it, end of story. Anyone with a competent design team can now compete with Intel with superior transistors and very little capital investment in comparison since the fab R&D and risk is shared across many LARGE customers.
Nothing in the near or medium term future can replicate smartphone leading edge volumes. PCs and servers are low single digit growth rates at best and still 10x lower volume by comparison. TSMC now has the economies of scale to stay ahead. You can’t compete against multiple trillion dollar customers funding R&D in Taiwan. The time to react was 12-15 years ago when the threat was still manageable. Now it’s out of reach.
Does our Board of Directors have the courage to pull the CEO if he goes astray?
Yes, Nvidia is a profit driven company making boatloads of money selling chips that drive new AI advances. And Intel is a basically non-profit that processes H1B applications for Desi family members from certain villages.
Nvidia and TSMC are currently leading. There is no argument with that. But Intel is holding strategic assets that other companies (Open AI and many cloud providers) are envy to have. Intel should leverage on its assets. Working towards AGI should be considered by Intel.
@oxj+1pFksPqR
Intel isn’t profit driven?
Maybe PG should take Intel private and re-incorporate as a non-profit.
Don’t quit your day job.
Why would he go to a sh1t tier company?
He’ll just do another startup.
AI hardware is not a moat for Intel.
You are very confused.
The moat belongs to Nvidia+TSMC which are literally 10x bigger than Intel.
Nvidia is very profit driven. I don't think it fits his value. Intel has the scope to work on AGI at a competitive cost base.
For humanity he should go to Nvidia, instead of Intel.