5 ms·

The thing is, though, that tying design and manufacturing together has been really good for Intel. Historically Intel's performance in terms of driver current

The thing is, though, that tying design and manufacturing together has been really good for Intel. Historically Intel's performance in terms of driver current and so forth at a given process node has been better than its competitors and it has reached those nodes more quickly. Intel manufacturing paid for that lead partially by accepting more restrictive design rules for transistor layout than a merchant silicon shop could ever get away with because they knew they only had one customer and one design they had to make work. With Intel's numerous missteps at 10nm that lead is over and the manufacturing debacle is taking down the design too but there's no reason to think Intel would be starting from such a dominant place if they had split up. If I were to go back to 2013 I'd tell them "10 nm is too early for cobalt interconnets!" rather than "Split up your design and manufacturing". And really you could also look at the problem as the design team being too wedded to x86 and unwilling to consider other ISAs just as much as the manufacturing problem. Intel could have done an ARM processor for the iPhone but they wanted to make a go at an x86 mobile chip instead. And while the x86 tax is small potatoes when you're paying for a deep OoO pipeline it's pretty noticeable in the world of in order processors or graphics processing units.

macintux 8y ago
That's exactly what Ben concludes with. > To demand that Intel apologize for its integrated model is satisfying in 2018, but all too dismissive of the 35 years of success and profits that preceded it. Many business failure stories are like that. "It worked great until it suddenly didn't."
- airstrike 8y ago
  IBM comes to mind
- pm90 8y ago
  Right. Its really hard to predict this kind of shifts in processes, or how processes that you're relied on for a long time have been made obsolete. When that happens, big businesses have an especially hard time because they're usually not agile enough to reorient themselves/adopt new processes quickly enough. A peer commenter mentioned IBM. As an ex-IBM'er, I certainly think that is true. The team I worked on adopted agile methodologies (which I think has been largely adopted in the emerging tech teams). But you could still see remnants of old processes and incentives, like a weird emphasis on writing patents. That was kinda shocking to me as a SE to see a patent writing as a part of my job, but I realized it was a holdover from earlier times. Ultimately, while I enjoyed my time there, broken processes for recognizing contribution was among the most important reason I jumped ship.
  - nickpsecurity 8y ago
    "That was kinda shocking to me as a SE to see a patent writing as a part of my job but I realized it was a holdover from earlier times." Useful activity for current times. Patent licensing and suits are a big money-maker and competition blocker for companies like IBM. For instance, Microsoft has made over a billion dollars on patent royalties *for Android." All that distracting paperwork paid off nicely. IBM used their legal team to crush a company selling commercial versions of Hercules emulator. Oracle vs Google. Apple vs Samsung. Intel recently threatened patent suits if anyone clones x86 to get off Intel chips. This stuff has a real, financial benefit to these companies. So, they're justified in investing in it.
    - pm90 8y ago
      You make an excellent point. I never thought of it that way and it totally makes sense. I guess I was looking at it from the perspective of an individual: I would rather my work be used to build amazing stuff than as way to extort others.
  - nostrademons 8y ago
    It's funny, though, that it's often very easy for an outsider to predict these kinds of shifts in processes, because they have no vested interest in the status quo and are evaluating everything with fresh eyes. As a kid growing up with computers in the 90's, I knew IBM was doomed. Why? Because buying an IBM-brand PC when a clone cost 1/2 the price and did exactly the same stuff was ridiculous, and Microsoft owned the software you actually used (and had to make sure all your apps were compatible with). Older folks would tell me "Well, IBM makes most of its money selling to enterprises, and they can charge millions of dollars for a computer" and I'd be like "Why would you do that when you can buy a PC for a couple thousand and hire a high-school kid to program it for you?" and they'd mutter something about transaction processing and customer support and how nobody got fired for buying IBM, and I'd be like "That's stupid. Their customers are all going to go bankrupt, and then IBM will too." I was a pretty insufferable teenager, but I was right about a lot of that. I'm 37 now, I've lived through four generations of technology (PC, web, mobile, and cloud) and retrained for each, and I find I need to make a conscious effort to try and see everything through college-student eyes. If I were a new grad just coming out of school, how would I evaluate the world? And I try to keep that in mind whenever I read eg. a thread on crypto here. The total volume of all data on the Ethereum blockchain is roughly what I would process in 10 minutes with a small MapReduce when I was working at Google. But it's very likely a college student today would be like "So? What's the point, when nobody trusts the answers you give them because you're just a giant corporation that keeps all your data secret?"
- Symmetry 8y ago
  It's possible, of course, that Intel is just doomed but I don't see that we have any reason to think Intel would be better off taking Ben's advice than not. Integration has hurt them in the last few years but again they could quite plausibly have ended up in a worse place now if they'd gone the foundry route years ago. I'm pretty certain that Intel's manufacturing arm would do much worse without being tied to the design arm and I think it's a tossup, going forward, whether the design arm will benefit or be hurt from being tied to the manufacturing arm.
  - justicezyx 8y ago
    Relentless reinvent itself, like Amazon did.
  - macintux 8y ago
    > but I don't see that we have any reason to think Intel would be better off taking Ben's advice than not Ben doesn't offer any advice that I can see, and he makes it clear that integration has worked very well for Intel up to this point. This seems like basic analysis, even though people seem determined to take it as advocacy.
- vinceguidry 8y ago
  This is what the moral of the story of Blockbuster was supposed to be, but everyone wanted a story of avoidable technological disruption, so they told themselves that one instead.
  - hype7 8y ago
    Ahhh, but that’s the thing. They’re the same story
- chasil 8y ago
  I disagree with a few of the points made by this article. "By 1986... Intel was already the best microprocessor design company in the world. They just needed to accept and embrace their destiny." ...they were most certainly not the best design. SPARC was just about ready to murder the VAX at this point, and more importantly, both the first MIPS and ARM samples were delivered in 1985. Either would have trounced an i386. The iPhone was also discussed. We should remember that Intel got StrongARM from Digital, later selling the Xscale architecture to Marvel. The destructive debris of ARM within Intel was likely substantial enough to have delivered whatever Steve Jobs wanted. However, Apple would not be happy with 32-bit ARM for long, and Intel should have realized that. If Intel had presented the equivalent of AArch64 to Apple a year before the iPhone, meeting power and code-density requirements with newly-designed 16- and 32-bit equivalents of THUMB within a 64-bit architecture, Apple might have abandoned ARM. Intel and HP shared Itanic - had Intel and Apple shared something like the A7, history would be quite different.
  - chasil 8y ago
    Reviewing the history, the Xscale/StrongARM sale to Marvel went through in June 2006. The original iPhone was released in June 2007 with a 620 MHz Samsung ARM. That was awfully close. I wonder if Apple was consulted on the sale, and what else was privately said between the two corporations. Intel certainly had the iPhone in their grasp. The Wiki says that the Blackberry Bold used an Xscale PXA930, the Torch used a PXA940, and the Xscale IOP348 was hitting 1.2GHz. Unbelievable that Intel walked away, after pushing several mobile products with the leading smartphone manufacturer of the time.
mtgx 8y ago
It's because integration and disintegration is all cyclical, as per Clayton Christensen (Innovator's Dilemma fame). When you have a new type of product you need integration to squeeze all of the performance out of it. When the industry is mature, you need to disintegrate it so benefit from outsourcing, economies of scale, customization features for customers, etc. Of course the devil is in the details, so you have to figure out when to do that exactly. Do it too early and it could be your downfall. Do it later, and the same could happen.
- scarface74 8y ago
  Jim Barksdale: “Gentlemen, there’s only two ways I know of to make money: bundling and unbundling.”

macintux 8y ago

That's exactly what Ben concludes with. > To demand that Intel apologize for its integrated model is satisfying in 2018, but all too dismissive of the 35 years of success and profits that preceded it. Many business failure stories are like that. "It worked great until it suddenly didn't."

airstrike 8y ago
IBM comes to mind
pm90 8y ago
Right. Its really hard to predict this kind of shifts in processes, or how processes that you're relied on for a long time have been made obsolete. When that happens, big businesses have an especially hard time because they're usually not agile enough to reorient themselves/adopt new processes quickly enough. A peer commenter mentioned IBM. As an ex-IBM'er, I certainly think that is true. The team I worked on adopted agile methodologies (which I think has been largely adopted in the emerging tech teams). But you could still see remnants of old processes and incentives, like a weird emphasis on writing patents. That was kinda shocking to me as a SE to see a patent writing as a part of my job, but I realized it was a holdover from earlier times. Ultimately, while I enjoyed my time there, broken processes for recognizing contribution was among the most important reason I jumped ship.
- nickpsecurity 8y ago
  "That was kinda shocking to me as a SE to see a patent writing as a part of my job but I realized it was a holdover from earlier times." Useful activity for current times. Patent licensing and suits are a big money-maker and competition blocker for companies like IBM. For instance, Microsoft has made over a billion dollars on patent royalties *for Android." All that distracting paperwork paid off nicely. IBM used their legal team to crush a company selling commercial versions of Hercules emulator. Oracle vs Google. Apple vs Samsung. Intel recently threatened patent suits if anyone clones x86 to get off Intel chips. This stuff has a real, financial benefit to these companies. So, they're justified in investing in it.
  - pm90 8y ago
    You make an excellent point. I never thought of it that way and it totally makes sense. I guess I was looking at it from the perspective of an individual: I would rather my work be used to build amazing stuff than as way to extort others.
- nostrademons 8y ago
  It's funny, though, that it's often very easy for an outsider to predict these kinds of shifts in processes, because they have no vested interest in the status quo and are evaluating everything with fresh eyes. As a kid growing up with computers in the 90's, I knew IBM was doomed. Why? Because buying an IBM-brand PC when a clone cost 1/2 the price and did exactly the same stuff was ridiculous, and Microsoft owned the software you actually used (and had to make sure all your apps were compatible with). Older folks would tell me "Well, IBM makes most of its money selling to enterprises, and they can charge millions of dollars for a computer" and I'd be like "Why would you do that when you can buy a PC for a couple thousand and hire a high-school kid to program it for you?" and they'd mutter something about transaction processing and customer support and how nobody got fired for buying IBM, and I'd be like "That's stupid. Their customers are all going to go bankrupt, and then IBM will too." I was a pretty insufferable teenager, but I was right about a lot of that. I'm 37 now, I've lived through four generations of technology (PC, web, mobile, and cloud) and retrained for each, and I find I need to make a conscious effort to try and see everything through college-student eyes. If I were a new grad just coming out of school, how would I evaluate the world? And I try to keep that in mind whenever I read eg. a thread on crypto here. The total volume of all data on the Ethereum blockchain is roughly what I would process in 10 minutes with a small MapReduce when I was working at Google. But it's very likely a college student today would be like "So? What's the point, when nobody trusts the answers you give them because you're just a giant corporation that keeps all your data secret?"
Symmetry 8y ago
It's possible, of course, that Intel is just doomed but I don't see that we have any reason to think Intel would be better off taking Ben's advice than not. Integration has hurt them in the last few years but again they could quite plausibly have ended up in a worse place now if they'd gone the foundry route years ago. I'm pretty certain that Intel's manufacturing arm would do much worse without being tied to the design arm and I think it's a tossup, going forward, whether the design arm will benefit or be hurt from being tied to the manufacturing arm.
- justicezyx 8y ago
  Relentless reinvent itself, like Amazon did.
- macintux 8y ago
  > but I don't see that we have any reason to think Intel would be better off taking Ben's advice than not Ben doesn't offer any advice that I can see, and he makes it clear that integration has worked very well for Intel up to this point. This seems like basic analysis, even though people seem determined to take it as advocacy.
vinceguidry 8y ago
This is what the moral of the story of Blockbuster was supposed to be, but everyone wanted a story of avoidable technological disruption, so they told themselves that one instead.
- hype7 8y ago
  Ahhh, but that’s the thing. They’re the same story
chasil 8y ago
I disagree with a few of the points made by this article. "By 1986... Intel was already the best microprocessor design company in the world. They just needed to accept and embrace their destiny." ...they were most certainly not the best design. SPARC was just about ready to murder the VAX at this point, and more importantly, both the first MIPS and ARM samples were delivered in 1985. Either would have trounced an i386. The iPhone was also discussed. We should remember that Intel got StrongARM from Digital, later selling the Xscale architecture to Marvel. The destructive debris of ARM within Intel was likely substantial enough to have delivered whatever Steve Jobs wanted. However, Apple would not be happy with 32-bit ARM for long, and Intel should have realized that. If Intel had presented the equivalent of AArch64 to Apple a year before the iPhone, meeting power and code-density requirements with newly-designed 16- and 32-bit equivalents of THUMB within a 64-bit architecture, Apple might have abandoned ARM. Intel and HP shared Itanic - had Intel and Apple shared something like the A7, history would be quite different.
- chasil 8y ago
  Reviewing the history, the Xscale/StrongARM sale to Marvel went through in June 2006. The original iPhone was released in June 2007 with a 620 MHz Samsung ARM. That was awfully close. I wonder if Apple was consulted on the sale, and what else was privately said between the two corporations. Intel certainly had the iPhone in their grasp. The Wiki says that the Blackberry Bold used an Xscale PXA930, the Torch used a PXA940, and the Xscale IOP348 was hitting 1.2GHz. Unbelievable that Intel walked away, after pushing several mobile products with the leading smartphone manufacturer of the time.

mtgx 8y ago

It's because integration and disintegration is all cyclical, as per Clayton Christensen (Innovator's Dilemma fame). When you have a new type of product you need integration to squeeze all of the performance out of it. When the industry is mature, you need to disintegrate it so benefit from outsourcing, economies of scale, customization features for customers, etc. Of course the devil is in the details, so you have to figure out when to do that exactly. Do it too early and it could be your downfall. Do it later, and the same could happen.

scarface74 8y ago
Jim Barksdale: “Gentlemen, there’s only two ways I know of to make money: bundling and unbundling.”