What about the Neural Engine? Will a new chip include tools for voice recognition and synthesis, and other ML tasks? That's a way of extending beyond Intel.
In a similar situation 40 years ago U.S. Steel bought an oil company and National Steel bought a drug distribution business.
So, what should Intel diversify into?
This.
With both Amazon and Apple moving to dump Intel in favor of in-house ARM designs fabbed by TSMC, the future does not look bright for Intel.
On the one hand, Intel totally deserves it -- this is the result of arrogance and complacency, pure and simple.
But on the other hand, I worry about this from a US national security perspective. I mean, good for TSMC -- they've accomplished a lot. But the PLA is a hop, skip, and a jump away.
I'd like to see Apple do some joint ventures with TSMC to build fabs further away from China. It doesn't even have to be the US -- put them in Japan or Germany -- but can we get a little distance from the red army please?
There's no way this thing doesn't run some version of ios. I realize we may not like it, but Apple currently has basically two product lines: one of them is synonymous with success and money and the other has been on some version of life support since its inception.
Especially on the low end, Apple sells MacBooks to college freshman who have their whole lives invested in their iPhone, but who use their Mac *exclusively* to "run" googledocs. And if a file ever does download to their harddrive, they have no idea how to find it.
If this thing runs iOS, it’ll never get my dollar. iOS is great for what it is but still sucks compared to Windows or MacOS when doing serious work. Just shifting files around on iOS is a nightmare.
The day MacOS goes, is the day I’ll buy my last MAC.
You seem to be putting a lot of stock in the mere names used for operating systems. Any well-architected software allows supporting multiple hardware architectures by moving pieces of the OS to different platforms. Apple can call it anything they like.
In a similar situation 40 years ago U.S. Steel bought an oil company and National Steel bought a drug distribution business.
So, what should Intel diversify into?
This.
With both Amazon and Apple moving to dump Intel in favor of in-house ARM designs fabbed by TSMC, the future does not look bright for Intel.
On the one hand, Intel totally deserves it -- this is the result of arrogance and complacency, pure and simple.
But on the other hand, I worry about this from a US national security perspective. I mean, good for TSMC -- they've accomplished a lot. But the PLA is a hop, skip, and a jump away.
I'd like to see Apple do some joint ventures with TSMC to build fabs further away from China. It doesn't even have to be the US -- put them in Japan or Germany -- but can we get a little distance from the red army please?
The US's manufacturing prowess has been relegated to the "more important" shareholder ROI. Since Jack Welch took over GE. We're now 40 years into the decline. That's two generations of money-men controlling the economy and economic and political and educational philosophy. Do we have two generations to rebuild the US back up again? I doubt it.
Same as what happened to Nokia and Blackberry when the iPhone came out. Intel’s history goes all the way back to the invention of the transistor. William Shockley’s engineers left to form Fairchild Semiconductor and later founded Intel. But if a trillion dollar company is unhappy with the performance of your products all the history in the world won’t save you, especially if you get outgunned by the competition.
It'll be very interesting to see the speed of conversion: Back in 2006, Apple announced the first Intel based Mac, and within 12 months had switched all of their products to Intel. But at least they provided Rosetta so you could still run PowerPC binaries. I sincerely hope they go the fat binary route again, so active software can still provide full performance on both ARM and Intel based Macs. (Just check that box in XCode and hit compile).
It is also important to do it as fast as possible, late 2020 is already a bit late. Maybe no one sees this coming but competition from opensource hardware and software designs will be intense.
Yeah - it's a shame - if Apple weren't so heavily invested in ARM they could've gone with RISC-V. I think that had it not been for AMD's fire storm of a comeback, x86 would already be on the way out. That's what prompted all the huge companies to create business server ARM CPUs in the first place. It would also be nice if Apple shifted to a non-binary delivery model to make ISA shifts less painful in the future.
What does this imply for Windows VMs and BootCamp? Like it or not, the business world is still dominated by Windows, and being able to run Windows is a must have for many people.
Apple dominates in the phone & tablet markets, and can support custom CPUs there, but do they really want to take on the desktop CPU world? Or are Intel & AMD saddled with lots of legacy cruft that Apple would be well-served to get away from? But I still fear becoming too divergent from the dominant desktop CPU architecture will end up hurting them.
Like I said before, if Apple makes an ARM core running x86 instructions, it minimize the transition problems for existing apps.
I'm not saying this is what Apple will do, but If I were Apple, this is what I would do, hear me out on this:
1, x86 instructions are the processor codes, not the OS. Both Windows and OS X are compiled into x86 instructions. 2, There used to be a RISC vs CISC architecture war to see which is faster. But Intel (and AMD) started to use x86 instructions on top of RISC-like cores and lines were blurred. 3, Intel just entered 10 nm process but Apple A14 is already going to use 5 nm, there may be room for x86 compatibility 4, Make the compatibility at the lowest level (closer to the hardware) helps existing software run faster. 5, Apple may need to license x86 instructions, AMD has a cross license deal with Intel on x86 for AMD's x86-64. 6, Intel may decide it is better to get a licensing deal with Apple since it is profit without the cost of manufacturing.
I'd like to see Apple do some joint ventures with TSMC to build fabs further away from China. It doesn't even have to be the US -- put them in Japan or Germany -- but can we get a little distance from the red army please?
TSMC is in Taiwan. It does have factories in China, but for lower end customers. It does have facilitates in the U.S.
It is also important to do it as fast as possible, late 2020 is already a bit late. Maybe no one sees this coming but competition from opensource hardware and software designs will be intense. Pine64 makes ARM hardware with Linux on it (not Android!) which is a decision as good as Apples macOS (Unix) on ARM. The point is that a Pine phone costs $150, and no it isn’t junk at all it is pretty impressive. On such hardware it is possible to install (for example) openbsd, one of the most secure and unhackable oses of this time,.. I’m doing that as a project on my Rock64. Not being in a closed system has a lot of important benefits.
Pine: no following, no software, no App Store, no nothing other than cheap. Hardly anything than another also ran. People buy Apple products because they work.
It’s interesting to read the comments. Reminds me of reading the news of Covid-19. How long will the virus last? How can you get it? Will it run windows? I’m not going to wear a mask! I’m not going to learn Xcode! What does this mean? Can I still go to the supermarket and not get the virus? Will they have some kind of emulator like Rosetta?
Truth is, it’s going to happen and once the wait is over things will eventually get back to normal. Just like the transition from Moto to Intel...
Question: if Windows has an ARM version of its OS, could an ARM based Mac Bootcamp or dual boot to ARM Windows? However bad the ARM Windows might be right now...
I'd like to see Apple do some joint ventures with TSMC to build fabs further away from China. It doesn't even have to be the US -- put them in Japan or Germany -- but can we get a little distance from the red army please?
TSMC is in Taiwan. It does have factories in China, but for lower end customers. It does have facilitates in the U.S.
I know. But China views Taiwan as a renegade province that they may choose to take by force whenever it suits them.
As time goes by, China's capacity to actually do that increases and the US' capacity/interest in stopping them decreases.
And then of course there's the potential influence of Chinese spies.
Bottom line --- it is not safe for the US to become dependent on chips fabbed in Taiwan.
Hopefully this will come with iPadOS and not MacOS.
What would be the point of that? If you have an iPad with a keyboard you've got the same thing.
Except that it's not the same thing. The iPad Pro with keyboard is thicker and heavier than the MacBook Air. And understandably so, because the keyboard case has to counter the heavier iPad. If the circuitry, casing and battery are in the base, Apple can make a lighter, thinner iPadOS product. I'd like to see how Apple is going to solve for hiding the keyboard. I believe this is why Adobe is bringing Illustrator and Photoshop to iPadOS.
It is also important to do it as fast as possible, late 2020 is already a bit late. Maybe no one sees this coming but competition from opensource hardware and software designs will be intense. Pine64 makes ARM hardware with Linux on it (not Android!) which is a decision as good as Apples macOS (Unix) on ARM. The point is that a Pine phone costs $150, and no it isn’t junk at all it is pretty impressive. On such hardware it is possible to install (for example) openbsd, one of the most secure and unhackable oses of this time,.. I’m doing that as a project on my Rock64. Not being in a closed system has a lot of important benefits.
Pine: no following, no software, no App Store, no nothing other than cheap. Hardly anything than another also ran. People buy Apple products because they work.
Looked it up. Just another iKnockoff using Apples technology. Thumbs up for the ounce of software originality though. When companies make anything original it's an complicated, ugly mess.
P.S. Apple Insider seems to be flooding the site with ARM Mac stories. Probably to secure future articles explaining how they predicted these. "In 2020 Apple Insider told you these were coming", with link.
Like I said before, if Apple makes an ARM core running x86 instructions, it minimize the transition problems for existing apps.
I'm not saying this is what Apple will do, but If I were Apple, this is what I would do, hear me out on this:
1, x86 instructions are the processor codes, not the OS. Both Windows and OS X are compiled into x86 instructions. 2, There used to be a RISC vs CISC architecture war to see which is faster. But Intel (and AMD) started to use x86 instructions on top of RISC-like cores and lines were blurred. 3, Intel just entered 10 nm process but Apple A14 is already going to use 5 nm, there may be room for x86 compatibility 4, Make the compatibility at the lowest level (closer to the hardware) helps existing software run faster. 5, Apple may need to license x86 instructions, AMD has a cross license deal with Intel on x86 for AMD's x86-64. 6, Intel may decide it is better to get a licensing deal with Apple since it is profit without the cost of manufacturing.
It won't happen because it's not necessary.
Unlike in previous architecture transitions, Apple not only has total control over the software stack (from kernel to APIs to development tools) but they have also totally removed all of the last vestiges of the classic MacOS. Everything is now not only based on NeXT technology, but that technology has gone though several generations of refinement. Even back in the 90s, NeXTStep was easily ported to multiple CPU architectures (68k; PPC; x86). Today, portability is even easier thanks to all the work Apple has put into things like LLVM, Clang, and Swift.
I predict that it will be very easy for the vast majority of developers to port to ARM -- far easier than prior transitions. The only need for x86 emulation would be apps whose developers have got out of business or chosen to abandon the apps, but by definition, those apps don't have much of a future anyway. Any app that is still being sold and updated will be quickly ported.
Apple ARM Macbook Pro; count me in. I have confident that when Apple releases ARM based MAC, performance will be if not better than on par with Intel/AMD offerings.Not only that but Apple will have long term plan for foreseeable future. .On software, long as major titles are ported, rest will follow fast.
Count me out. Projects like this are a complete waste of resources.
Comments
With both Amazon and Apple moving to dump Intel in favor of in-house ARM designs fabbed by TSMC, the future does not look bright for Intel.
On the one hand, Intel totally deserves it -- this is the result of arrogance and complacency, pure and simple.
But on the other hand, I worry about this from a US national security perspective. I mean, good for TSMC -- they've accomplished a lot. But the PLA is a hop, skip, and a jump away.
I'd like to see Apple do some joint ventures with TSMC to build fabs further away from China. It doesn't even have to be the US -- put them in Japan or Germany -- but can we get a little distance from the red army please?
It'll be great to have 32 registers again.
Apple dominates in the phone & tablet markets, and can support custom CPUs there, but do they really want to take on the desktop CPU world? Or are Intel & AMD saddled with lots of legacy cruft that Apple would be well-served to get away from? But I still fear becoming too divergent from the dominant desktop CPU architecture will end up hurting them.
I'm not saying this is what Apple will do, but If I were Apple, this is what I would do, hear me out on this:
1, x86 instructions are the processor codes, not the OS. Both Windows and OS X are compiled into x86 instructions.
2, There used to be a RISC vs CISC architecture war to see which is faster. But Intel (and AMD) started to use x86 instructions on top of RISC-like cores and lines were blurred.
3, Intel just entered 10 nm process but Apple A14 is already going to use 5 nm, there may be room for x86 compatibility
4, Make the compatibility at the lowest level (closer to the hardware) helps existing software run faster.
5, Apple may need to license x86 instructions, AMD has a cross license deal with Intel on x86 for AMD's x86-64.
6, Intel may decide it is better to get a licensing deal with Apple since it is profit without the cost of manufacturing.
As time goes by, China's capacity to actually do that increases and the US' capacity/interest in stopping them decreases.
And then of course there's the potential influence of Chinese spies.
Bottom line --- it is not safe for the US to become dependent on chips fabbed in Taiwan.
Looked it up. Just another iKnockoff using Apples technology. Thumbs up for the ounce of software originality though. When companies make anything original it's an complicated, ugly mess.
P.S. Apple Insider seems to be flooding the site with ARM Mac stories. Probably to secure future articles explaining how they predicted these. "In 2020 Apple Insider told you these were coming", with link.
It won't happen because it's not necessary.
Unlike in previous architecture transitions, Apple not only has total control over the software stack (from kernel to APIs to development tools) but they have also totally removed all of the last vestiges of the classic MacOS. Everything is now not only based on NeXT technology, but that technology has gone though several generations of refinement. Even back in the 90s, NeXTStep was easily ported to multiple CPU architectures (68k; PPC; x86). Today, portability is even easier thanks to all the work Apple has put into things like LLVM, Clang, and Swift.
I predict that it will be very easy for the vast majority of developers to port to ARM -- far easier than prior transitions. The only need for x86 emulation would be apps whose developers have got out of business or chosen to abandon the apps, but by definition, those apps don't have much of a future anyway. Any app that is still being sold and updated will be quickly ported.