4 ms·

Apple has already won. Even if Intel catches up in efficiency, it doesn't matter. Even if they're 30% faster, it doesn't matter. Apple is in charge of their own

Apple has already won. Even if Intel catches up in efficiency, it doesn't matter. Even if they're 30% faster, it doesn't matter. Apple is in charge of their own destiny with Apple Silicon. That's invaluable.

abledon 5y ago
isnt it TSMC's silicon with a printing recipe sent over from apple?
- collinvandyck76 5y ago
  Yes, Apple relies on other foundries to build their chips.
- eyelidlessness 5y ago
  Well, that’s a surprisingly myopic take on a site with a disproportionately large audience who send recipes to silicon on the regular.
- rstupek 5y ago
  Doesn't AMD similarly send their recipe to TSMC to be printed?
  - lostmsu 5y ago
    I think TSMC being the king here was the point of the comment.
- pjscott 5y ago
  That’s correct, but (a) the recipe is extremely non-trivial and (b) Apple could take that recipe to Samsung and get mostly comparable results for something in the double-digit millions of dollars range for design conversion, so they have leverage.
howdydoo 5y ago
So Apple is the best regardless of which product is better?
- reaperducer 5y ago
  I think he's saying that Apple processors are best for Apple. It's no longer beholden to outsiders for its processors. There is value in freedom that, for Apple, is greater than mere numbers can convey.
- CharlesW 5y ago
  I think it depends on one's definition of "better". There will likely always be performance-at-any-cost choices that will beat Apple since that's not what they're solving for.
- jimbob45 5y ago
  I think that’s what he’s trying to say? It doesn’t make sense though because having your own architecture isn’t that useful if no one else wants to use it. The goal is to 10X your profits by licensing out your designs. Designing your own architecture is a 1X move if only you use it and that’s not counting the cost of maintenance and marketing. Counterpoint: Apple has sunk an incredible amount of money into an architecture that provides very few tangible benefits to the consumer. It looks unlikely that they’ll ever recoup costs. They could have put that money toward their own fab and received infinite US government money instead.
  - dasil003 5y ago
    What you are saying is antithetical to the entire Apple strategy, which is to create the best vertically integrated hardware and software and charge a premium for it. The strategy you're suggesting is to commoditize yourself, which they already tried (do you remember the Mac clone era), and is also the same reason IBM lost the PC market to Microsoft and none of the resulting hardware OEMs have even a fraction of Apple's market cap. Also, to suggest that Apple Silicon has very few tangible benefits is laughable. The power / performance are incredible, and the deep integration provides tons of optimizations that no other manufacturer can match. This ranges from seemingly simple things like M1 Macs waking up faster than you can open the lid, to long-term strategic advantages such as cross-compiling single binaries to both Mac OS and iOS, or adding custom silicon to cover software features with unprecedented efficiency with a guarantee that every single machine running it will have access to that hardware (over time).
jayd16 5y ago
Maybe not at 10% but if Intel was 30% faster, I think it would raise some eyebrows.
howaboutnope 5y ago
Apple Silicon comes with Apple software. It's worthless to me.
jonstewart 5y ago
PowerPC would like a word.
touchpadder 5y ago
"Even if they're 30% faster, it doesn't matter" of course it doesn't lol
hakfoo 5y ago
Every time they zig or zag on the architecture, though, it presents a risk to their third-party ecosystem. Yeah, there will be a few years where a backwards-compatibility penalty box is offered, but eventually someone's x86-64 only tools or extensions will break. They also threw out the ability to say "If your company/tasks don't work with MacOS, you can always dual-boot full-fat Windows." They have to be betting that their platform is still big and compelling enough to keep bringing in users and developers to replace anything they lose to the x86-64 abandonment. I tend to think that there's no real magic secret to ARM for low power, it's more that we've spent the last 20 years optimizing them for low-power devices while x86 was traditionally performance-at-all-cost.

abledon 5y ago

isnt it TSMC's silicon with a printing recipe sent over from apple?

collinvandyck76 5y ago
Yes, Apple relies on other foundries to build their chips.
eyelidlessness 5y ago
Well, that’s a surprisingly myopic take on a site with a disproportionately large audience who send recipes to silicon on the regular.
rstupek 5y ago
Doesn't AMD similarly send their recipe to TSMC to be printed?
- lostmsu 5y ago
  I think TSMC being the king here was the point of the comment.
pjscott 5y ago
That’s correct, but (a) the recipe is extremely non-trivial and (b) Apple could take that recipe to Samsung and get mostly comparable results for something in the double-digit millions of dollars range for design conversion, so they have leverage.

howdydoo 5y ago

So Apple is the best regardless of which product is better?

reaperducer 5y ago
I think he's saying that Apple processors are best for Apple. It's no longer beholden to outsiders for its processors. There is value in freedom that, for Apple, is greater than mere numbers can convey.
CharlesW 5y ago
I think it depends on one's definition of "better". There will likely always be performance-at-any-cost choices that will beat Apple since that's not what they're solving for.
jimbob45 5y ago
I think that’s what he’s trying to say? It doesn’t make sense though because having your own architecture isn’t that useful if no one else wants to use it. The goal is to 10X your profits by licensing out your designs. Designing your own architecture is a 1X move if only you use it and that’s not counting the cost of maintenance and marketing. Counterpoint: Apple has sunk an incredible amount of money into an architecture that provides very few tangible benefits to the consumer. It looks unlikely that they’ll ever recoup costs. They could have put that money toward their own fab and received infinite US government money instead.
- dasil003 5y ago
  What you are saying is antithetical to the entire Apple strategy, which is to create the best vertically integrated hardware and software and charge a premium for it. The strategy you're suggesting is to commoditize yourself, which they already tried (do you remember the Mac clone era), and is also the same reason IBM lost the PC market to Microsoft and none of the resulting hardware OEMs have even a fraction of Apple's market cap. Also, to suggest that Apple Silicon has very few tangible benefits is laughable. The power / performance are incredible, and the deep integration provides tons of optimizations that no other manufacturer can match. This ranges from seemingly simple things like M1 Macs waking up faster than you can open the lid, to long-term strategic advantages such as cross-compiling single binaries to both Mac OS and iOS, or adding custom silicon to cover software features with unprecedented efficiency with a guarantee that every single machine running it will have access to that hardware (over time).

jayd16 5y ago

Maybe not at 10% but if Intel was 30% faster, I think it would raise some eyebrows.

howaboutnope 5y ago

Apple Silicon comes with Apple software. It's worthless to me.

jonstewart 5y ago

PowerPC would like a word.

touchpadder 5y ago

"Even if they're 30% faster, it doesn't matter" of course it doesn't lol

hakfoo 5y ago

Every time they zig or zag on the architecture, though, it presents a risk to their third-party ecosystem. Yeah, there will be a few years where a backwards-compatibility penalty box is offered, but eventually someone's x86-64 only tools or extensions will break. They also threw out the ability to say "If your company/tasks don't work with MacOS, you can always dual-boot full-fat Windows." They have to be betting that their platform is still big and compelling enough to keep bringing in users and developers to replace anything they lose to the x86-64 abandonment. I tend to think that there's no real magic secret to ARM for low power, it's more that we've spent the last 20 years optimizing them for low-power devices while x86 was traditionally performance-at-all-cost.