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

Intel Honesty

stratechery.com

Post reply on HN
Print view View original

81–90 of 426 posts

Re: Intel Honesty

#81

Post by amluto »

andy_xor_andrew wrote: post #6

US chip production really needs its SpaceX moment, but it seems like it will never happen, and we're left with the crumbling empire of Intel. By "SpaceX moment", I mean a startup entering an impossible market, where the barrier to entry is billions of dollars of research and manufacturing, a market dominated by industry giants from the 60s, and yet still coming out on top somehow.

Maybe one of the e-beam startups can pull it off. Aside from scaling issues, direct electron beam lithography ought to outperform optical lithography by many metrics, not to mention that there would never be costs associated with mask revisions.

Here’s one of them: https://multibeamcorp.com/

I remember touring a little research fab, maybe around 2000, that could achieve feature sizes comparable to what TSMC can do today. But they were very, very, very slow.

(Fast-moving electrons are easy to make, easy to aim, and have teeny tiny wavelengths that entirely sidestep most the issues that people have with photons having obnoxiously large wavelengths. But electrons have all manner of downsides that explain why fabs spend many billions of dollars on optical lithography, one of which is that they repel each other, which makes shining a lot of them at a wafer at once quite problematic.)

Re: Intel Honesty

#82

Post by zhobbs »

api wrote: post #20
scrlk wrote: post #10

> in the late 2010’s Intel got stuck trying to move to 10nm, thanks in part to their reluctance to embrace the vastly more expensive EUV lithography process TBH, it's easy to say this with the benefit of hindsight. Throughout most of the 2010s, EUV lithography was like the "year of the Linux desktop" - i.e., this year will be the year where EUV was suitable for high volume manufacturing. I don't really blame Intel fo…

> Throughout most of the 2010s, EUV lithography was like the "year of the Linux desktop" - i.e., this year will be the year where EUV was suitable for high volume manufacturing. We need to learn to recognize the difference between something that's never going to happen for either physical or economic/social/structural reasons, and something that is just really difficult and takes a long time. I always think of this w…

It’s a good point, but one thing to consider is that ultra long hard tech might be structurally challenging as well. If fusion requires 100 years of capital and R&D it won’t be viable ever due to economic/societal/structural reasons.

Re: Intel Honesty

#83

Post by NortySpock »

Invictus0 wrote: post #74
bryanlarsen wrote: post #49

Earlier quoted context omitted.

The payments are tied to real milestones like building fabs. Intel cannot spend the money on organizational inertia because that wouldn't get them the money.

It doesn't matter how badly they want to sell the chips if they don't have the money to build the fabs in the first place. That's where Intel is at right now. Building fabs is incredibly expensive and they need the money upfront.

You can get a loan against the value of a purchase contract if you have a firm purchase contract (i.e. from the government) and a convincing plan on how to manufacture to that price point...

Re: Intel Honesty

#84

Post by cs702 »

The OP makes a compelling case that Intel's foundry operations have fallen so far behind that they are no longer economically viable on their own, in the face of current market forces, i.e., Intel's foundry business may not be able to earn a positive return on the tens of billions of dollars of capex now required to catch up with TSMC.

If the US truly views having a domestic foundry as critical to national security, the US federal government has no choice but to pay up big time to support Intel's foundry business until -- hopefully, eventually -- that business is able to compete profitably against TSMC to manufacture chips for Apple, Nvidia, AMD, etc.

Oh, how the mighty have fallen!

Re: Intel Honesty

#85

Post by amluto »

ska wrote: post #67
highfrequency wrote: post #61

Earlier quoted context omitted.

Nice analysis on competitive dynamics of semiconductor industry vs. space industry in the 2000s. > I'm not aware of any other industries which are quite as far "behind" today as space launch systems were in 2010. But if anyone else is, please mention those industries here! Digital payments may be in a similar situation. Visa's 2% fee may seem low but from a scale and competitive standpoint it is fairly absurd - they…

From what i can see as an admitted outsider, payment rails are more complicated than they look, and most of the crypto partisans arguments to "replace" functionality are a mix of 'ignoring that part', or 'we don't want that part anyway' + a few API calls. Problem is, the world consuming these things mostly seems to want those parts, and can't ignore those other parts. Which seems to explain relatively low uptake.

Those rails involve an amazing amount of complexity that boils down to the entire system being nonsensical.

Payments are pulled instead of pushed; the underlying credit card numbers lack even a semblance of security; there is all kinds of mis-design due to the way that restaurant tips work; it all started when credit card imprints/readers were all assumed to be offline; etc.

Re: Intel Honesty

#86

Post by ho_schi »

SoftTalker wrote: post #28
api wrote: post #20

Earlier quoted context omitted.

> Throughout most of the 2010s, EUV lithography was like the "year of the Linux desktop" - i.e., this year will be the year where EUV was suitable for high volume manufacturing. We need to learn to recognize the difference between something that's never going to happen for either physical or economic/social/structural reasons, and something that is just really difficult and takes a long time. I always think of this w…

I'd say Linux desktops have been good enough for a long time, but never broke into the mainstream for the reasons you cite. ChromeOS is the closest thing I guess, but outside of schools and a few other institutional uses, Chromebooks are not very popular.

Linux is not preinstalled. There are some ThinkPads* and Dells with Ubuntu or Fedora but you need to know that.

And people how know will and must reinstall anyway Arch, Gentoo, Suse, Debian, Fedora or Ubuntu.

The Steamdeck is an excellent proof how a full featured Linux (SteamOS is based upon Arch) is shipped well. Not a unmaintained or googlyified closed-source derivate of Linux (Android and ChromeOS).

* I ordered an X13 Gen1 AMD with Linux for fun. Worked well, installation was clean. No 120$ for Microsoft and its stock owners. You shall not feed the bad guys. Especially when you will never use Windows.

Re: Intel Honesty

#87

Post by numpad0 »

andy_xor_andrew wrote: post #6

US chip production really needs its SpaceX moment, but it seems like it will never happen, and we're left with the crumbling empire of Intel. By "SpaceX moment", I mean a startup entering an impossible market, where the barrier to entry is billions of dollars of research and manufacturing, a market dominated by industry giants from the 60s, and yet still coming out on top somehow.

I think lack of focus on US domestic chip capability is semi-intentional and is an eventuality. Semiconductor industry, and many other manufacturing industries too, seem to flourish at very outer edges of the free world.

It could be simple as the free world not wanting manufacturing capability at home. It could be wanting only grain farms and money printers.

Re: Intel Honesty

#88

Post by reducesuffering »

ryandrake wrote: post #70
psunavy03 wrote: post #36

Earlier quoted context omitted.

I don't get how some people struggle with the idea that he may simultaneously be a gifted businessman (in certain sectors, obviously Twitter was a disaster) and also simultaneously be a raging asshole. Bill Gates was a raging asshole. Steve Jobs was a raging asshole. Larry Ellison was a raging asshole. It's not ideal, but it also doesn't preclude success.

It isn't required for success, either. If Elon Musk was never born, it is likely we'd still have SpaceX and Tesla equivalents, but with a better human being at the helm. A parallel universe also exists where Microsoft, Apple, and Oracle were not founded by raging assholes. The fact that these companies are all run by jerks is a terrible coincidence, but it does not necessarily need to be so.

> If Elon Musk was never born, it is likely we'd still have SpaceX and Tesla equivalents

I'm perturbed by Elon's politics and hypocrisy, but this is absolutely not true in anywhere near the same timeframe.

Re: Intel Honesty

#89

Post by schmidtleonard »

hangonhn wrote: post #14
scrlk wrote: post #10

> in the late 2010’s Intel got stuck trying to move to 10nm, thanks in part to their reluctance to embrace the vastly more expensive EUV lithography process TBH, it's easy to say this with the benefit of hindsight. Throughout most of the 2010s, EUV lithography was like the "year of the Linux desktop" - i.e., this year will be the year where EUV was suitable for high volume manufacturing. I don't really blame Intel fo…

In addition, SMIC in China managed to get to 7 nm without the use of EUV. Also early EUV yields were actually quite low.

Did SMIC get 7nm to yield or is it limping but propped up for PR purposes like Intel 7?

Re: Intel Honesty

#90

Post by jhallenworld »

The Altera merger pain is maybe enlightening.

Altera 10-series FPGAs were massively delayed due to Intel's 10 nm fab problems. But I speculate that there was also a big difference in toolchains, does anyone know for sure? I mean Altera was using TSMC previously, and I presume were using industry standard tools: Cadance / Synopsis. But I would guess Intel was running their fab on home-grown tools... what is the status these days? For example I know for sure that IBM's synthesis was "Booledozer".

It's interesting because maybe Intel will spin out the fab business. But are they immediately ready to become a commercial fab business instead an Intel-only business? What's the toolchain like for Intel fab?

Post reply on HN