Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

US Intel

stratechery.com

Post reply on HN
Print view View original

241–250 of 581 posts

Re: US Intel

#241

Post by Aurornis »

jollyllama wrote: post #232
georgeburdell wrote: post #23

Earlier quoted context omitted.

If I may add my view as a formerly high-achieving semiconductor worker that Intel would benefit greatly from having right now, a lot of us pivoted to software and machine learning to earn more money. My first 2 years as a software engineer earned me more RSUs than a decade in semiconductors. Semiconductors is not prestigious work in the U.S., despite the strategic importance. By contrast, it is highly respected and r…

Billions for fast food delivery apps, not a dime for defense or sustainable agriculture.

> Billions for fast food delivery apps, not a dime for defense or sustainable agriculture

Defense venture capital funding is literally in the billions and has resilient even in a broader VC pullback: https://www.spglobal.com/market-intelligence/en/news-insight...

Re: US Intel

#242

Post by chatmasta »

nemothekid wrote: post #235
ecshafer wrote: post #195

Earlier quoted context omitted.

How many workers do we need? TSMC has 80k workers all in. 15-24 there is about 3 million US NEETs. I think that we could have a 10 year plan to get 80k people with EE/CE/NanoTech Eng/ChemEng/etc degrees. Make college free, or hell even a stipend for people to pursue "Critical Careers", heavily fund Phds in the same areas for US Citizens, and we will see workers with the skills.

This is the wrong way to look at it. The people in the US that you would want working in semiconductors aren't NEETs. They are at Facebook/Google/Jane Street/Citadel/McKinsey. If I am a capable person working on delivering node improvements dealing with smaller and smaller challenges as the physics issues become quantum - I will eventually start to ask myself: why am I working on the hardest physics problems in the p…

Is that true? I thought a lot of semi-conductor work is borderline blue-collar factory work and physical labor.

What you’re describing is in the R&D area and also not physically dependent on being colocated in a fab. So we should have an easier time finding that talent, although we’re probably underpaying them now, as you point out.

Re: US Intel

#243

Post by onetimeusename »

bigyabai wrote: post #225
onetimeusename wrote: post #220

Earlier quoted context omitted.

I'm not sure they need to claim Intel as eminent domain and that type of move is beyond what I know about but I suppose they could've done that previously but opted instead for market based solutions like with the CHIPS Act. That is really what led to the equity stake. The purpose of the CHIPS Act had more to do with the AI Cold War and securing the supply chain than it did trying to save one particular business. So…

> The purpose of the CHIPS Act had more to do with the AI Cold War and securing the supply chain than it did trying to save one particular business I just don't think we're ever going to see eye-to-eye if this is your belief. In realpolitiks terms, Intel hasn't been a player on the AI board since Gaudi. And even that was a total flop. If securing AI compute was the goal, buying Intel is about as large of a mistake as…

I think it's simpler than that. I think US intelligence is looking to make sure a secure supply chain exists for AI and really that just means NVDA's chips. I have never stated at any point this plan would work well. This is all stated in the reasoning for the CHIPS Act. That may mean pressuring people to use the foundry but I don't actually think the concern was chiefly to win back jobs. The funding from the CHIPS Act was actually withheld since Intel wasn't making satisfactory progress. Also the warrants I mentioned are executable only if Intel drops it's stake in the foundry. I believe this was added since there was talk of Intel spinning that out. So the government seems very interested in the viability of the foundry and less so in jobs.

So my point is that I think US Defense is motivated by national security interests but that doesn't mean this plan is going to work well or that we haven't already lost the AI war. It's probably too little too late. I am just pointing out over and over that I don't think grandstanding is the motive. I am sure that will happen but I think at some point it hit US Defense that relying on Taiwan put's us in a precarious situation.

Re: US Intel

#244

Post by etempleton »

webdevver wrote: post #60

for all this talk about intel struggling, they seem to be quite capable to continue producing very competitive CPUs and other IP. my understanding is that they are still doing well in the datacentre world, unless that has changed? its worth keeping in mind Intel is quite a big company... and naturally, the parts that are chugging along just fine are not going to be making headlines (or much noise at all really.)

There is a lot of doom and gloom regarding Intel, but there are also signs of life. Intel is starting to show off products made on 18a and is slated to ship at least one product on that process this calendar year. Meanwhile TSMC has been stuck on 3nm for quite some time and it appears that there will not be a product with TSMC 2nm this year. If both of these prove true, Intel will have roughly caught up in terms of process. There is a lot more work to be done, but do not underestimate how quickly Intel's fortune can change if they manage to start hitting their mark and TSMC falters even slightly.

Re: US Intel

#245

Post by snapcaster »

RealityVoid wrote: post #130
makestuff wrote: post #126

Earlier quoted context omitted.

Yeah seeing the innovations DJI is making in agriculture makes me wonder why VCs do not want to fund other startups like it. I know the margins are worse and it takes way more capital to fund hardware companies, but that has to be better than funding another chat-gpt wrapper. I guess that is why I am not a VC though lol.

They're just hard. What is the US robotics startup that you would use as an example of success? iRobot? It barely broke even.

Anduril?

Re: US Intel

#246

Post by Aurornis »

georgeburdell wrote: post #23
themgt wrote: post #7

I’ll be honest: there is a very good chance this won’t work .... At the same time, the China concerns are real, Intel Foundry needs a guarantee of existence to even court customers, and there really is no coming back from an exit. There won’t be a startup to fill Intel’s place. The U.S. will be completely dependent on foreign companies for the most important products on earth, and while everything may seem fine for t…

If I may add my view as a formerly high-achieving semiconductor worker that Intel would benefit greatly from having right now, a lot of us pivoted to software and machine learning to earn more money. My first 2 years as a software engineer earned me more RSUs than a decade in semiconductors. Semiconductors is not prestigious work in the U.S., despite the strategic importance. By contrast, it is highly respected and r…

The gap between hardware and software salaries has been strangely stubborn.

I started in hardware and pivoted to software for the same reason: It's easier to find higher paying jobs.

It's been strange to see the gap persist at companies that cling to salary data for compensation decisions. Incredible to see companies complain about not being able to find good hardware talent but then also refuse to pay hardware hires at the same level as software hires.

Re: US Intel

#247

Post by ecshafer »

nemothekid wrote: post #235
ecshafer wrote: post #195

Earlier quoted context omitted.

How many workers do we need? TSMC has 80k workers all in. 15-24 there is about 3 million US NEETs. I think that we could have a 10 year plan to get 80k people with EE/CE/NanoTech Eng/ChemEng/etc degrees. Make college free, or hell even a stipend for people to pursue "Critical Careers", heavily fund Phds in the same areas for US Citizens, and we will see workers with the skills.

This is the wrong way to look at it. The people in the US that you would want working in semiconductors aren't NEETs. They are at Facebook/Google/Jane Street/Citadel/McKinsey. If I am a capable person working on delivering node improvements dealing with smaller and smaller challenges as the physics issues become quantum - I will eventually start to ask myself: why am I working on the hardest physics problems in the p…

The cost of assets (especially housing, schooling, and health care) is a huge problem, and your example is a poignant one. More funding could sway a few people in that pipeline to go towards semi-conductors, but the majority of workers aren't Jane Street quality, they are technicians and engineers doing lots of highly skilled "grunt" work.

Personally I think the other half of the problem, Big Tech paying so much might be solving itself right now, excepting really only the very very top.

Re: US Intel

#248

Post by mrkstu »

nemothekid wrote: post #235
ecshafer wrote: post #195

Earlier quoted context omitted.

How many workers do we need? TSMC has 80k workers all in. 15-24 there is about 3 million US NEETs. I think that we could have a 10 year plan to get 80k people with EE/CE/NanoTech Eng/ChemEng/etc degrees. Make college free, or hell even a stipend for people to pursue "Critical Careers", heavily fund Phds in the same areas for US Citizens, and we will see workers with the skills.

This is the wrong way to look at it. The people in the US that you would want working in semiconductors aren't NEETs. They are at Facebook/Google/Jane Street/Citadel/McKinsey. If I am a capable person working on delivering node improvements dealing with smaller and smaller challenges as the physics issues become quantum - I will eventually start to ask myself: why am I working on the hardest physics problems in the p…

Look at the insane salaries/equity going to AI researchers. Those at the cutting edge of semi manufacturing/design are likely completely capable of reproducing the skills sets of those people- how do they not cut and run?

If the government has to provide the funds, so be it, make those jobs valuable enough and the skills will be there.

Re: US Intel

#249

Post by hedgehog »

scarface_74 wrote: post #205
onetimeusename wrote: post #189

The other day when the US's stake in Intel was announced, people assumed it was a political stunt. I suspected it was because of national security interests. The CHIPS act probably didn't get the result US Defense wanted quickly enough. Some details that were glossed over include that there was a chip shortage a few years ago as a result of COVID and TSMC supply chain disruptions that led to a shortage in electronics…

The issue is that Intel manufacturing chips in the US wouldn’t have solved the chip problems with cars for instance. What TSMC traditionally does is keep trailing edge fabs online that are fully depreciated and use those to produce chips that don’t need to be leading edge. It wouldn’t make sense to create a new fab for trailing edge chips. Car manufacturers aren’t going to all of the sudden start using 2mm expensive…

14nm is mature and would suit this purpose fine, and they have existing capacity. The business, tools, and everything else around would need to get built out.

Re: US Intel

#250

Post by ac29 »

JustExAWS wrote: post #145
qwytw wrote: post #131

Earlier quoted context omitted.

> This is demonstrably not true. In what way? TSMC doesen't design or sell the chips. If they have limited capacity they will of course charge more for manufacturing mobile chips if they can sell the capacity to Nvidia/AMD/Apple instead. ARM chips (and that's pretty much by design based on ARM's business model) are close to being a commodity. Apple is of course an exception but they are not directly part of the CPU m…

> TSMC doesen't design or sell the chips. If they have limited capacity they will of course charge more for manufacturing mobile chips if they can sell the capacity to Nvidia/AMD/Apple instead That’s the entire point. Intel is behind because their vertical market strategy of using their own fabs only for their own chips doesn’t give them enough volume compared to TSMC who has volume because they are a foundry. > ARM…

> every mobile cellular chip that I’m aware of except for the very few designed by Apple for the low end 16e is sold by Qualcomm

Not sure on volume, but Samsung and Mediatek are major players too. I dont think Google has used anything from Qualcomm since Pixel 5 (2020).

Post reply on HN