A 'post-x86 world'? Preposterous!

I honestly don't know whether Om Malik's blog site, GigaOM, is intended to be informative or merely entertaining. I pointed out a previous example of the overwrought rhetoric that permeates that site last September (in the context of Comcast's then-new usage cap policy), but generally, I try to ignore the nonsense there for the same reasons that I ignore talk radio.

But like it or not, GigaOM is widely read, and sometimes when a post there bears directly on a market that's important to me, I can't bear to let it go. This is one of those times.

On Thursday, a GigaOM staffer wrote a piece titled "Can Intel Thrive in a Post x86 World?"

A slide from Fred Weber's keynote presentation at Microprocessor Forum 2003 showing how x86 will evolve into systems from big servers down to handheld consumer devices.

(Credit: Advanced Micro Devices, Inc.)

The headline is preposterous from beginning to end. It has two implications just in the eight words of the title: that Intel's ability to "thrive" faces any imminent threats, and that the importance of the x86 architecture is declining.

In January, the same staffer wrote a piece titled "Netbooks and the Death of x86 Computing" which reached the fantastic conclusion that Netbooks would "destroy the hegemony of x86 machines for personal computing."

Well, as I pointed out just a few weeks later (in "The Netbook is dead. Long live the notebook!"), when the Netbook phenomenon ran up against the dominance of Intel and Microsoft in the PC market, it was the Netbook that died instead. Even at a $300 price point, people still want full PC compatibility.

Yes, there are companies like Freescale (the subject of the January post on GigaOM) and Nvidia that are looking to push the ARM architecture into the Netbook space. But that idea never made much sense, and now that Intel and TSMC are working together to get Intel's Atom x86 core into lower-cost SoC (system on chip) products, the ARM architecture will eventually have to retreat into the shrinking niche for supersmall, supercheap phones and consumer electronics gizmos for which x86 compatibility is of negligible value.

See, we learned a long time ago--those of us who cover this industry professionally, not just as a random assignment for some random blog--that the instruction set architecture (ISA), per se, doesn't matter any more.

The choice of ISA was a big deal in the 1980s and early 1990s, when the extra complexity of an x86 instruction decoder was a large fraction of the total complexity of a microprocessor. That's where the conflict between RISC and CISC came from.

But by the turn of the century, ISA complexity was almost a dead issue, and that coffin's final nail was pounded in by the keynote speech of then-Advanced Micro Devices CTO Fred Weber at Microprocessor Forum 2003, an event I had the honor of hosting.

In his talk, "Towards Instruction Set Consolidation," Weber made a simple point: "Technology has passed the point where instruction set costs are at all relevant."

Even then, three generations of process technology ago, the "x86 penalty" was down to a couple square millimeters of silicon. Today, the comparable figure is about 0.25 square millimeters. Not zero, certainly, but not a significant concern for chips that are a hundred times larger.

In short, ARM chips aren't cheaper or more power-efficient because of their instruction sets; they're like that because they're designed to be. And anything that an ARM chip can do to save cost or power can also be done by an x86 chip.

So there can't ever be a time when the world moves beyond x86. That's 1980s thinking, just plain ignorance of what may be the most important trend in the microprocessor industry.

The rest of Thursday's GigaOM post is a hopelessly self-contradictory muddle that fails to reach any clear conclusions. I'll just quote one more line near the end: "But the PC will be just one small (and shrinking) battleground to keep x86 relevant, amid a more mobile, visual, and power-sensitive world."

Current economic woes aside, the PC market is hardly shrinking. You know what's shrinking? The PC! As the PC shrinks, the PC market will grow. The MID (mobile Internet device) market isn't much to speak of right now, for example, but once MID makers figure out what to build, MIDs will become more popular.

And seriously, is anyone really not clear on the fact that the Apple iPhone is a computer? It isn't an embedded system. An embedded system is one in which the presence of a microprocessor is functionally irrelevant to the user. When a gizmo exposes its programmability to the user, it's a computer.

What else is the App Store but the visible manifestation of the iPhone's programmability?

Now, ARM isn't dead yet. The iPhone uses an ARM processor because there's no x86 processor that would work as well in that system. ARM processors will probably see at least two more generations in cell phones just because there's so much ARM-based software out there (including all the software on the App Store).

But somewhere around 2012, we're going to see x86 chips poking into that space. The value of instruction set compatibility with the PC market will persuade developers of new cell phone platforms to go with x86 chips, and eventually even established systems like the iPhone will switch over.

So not only are x86 chips selling into a growing PC market, they'll eventually start eating into ARM's own strongholds. That can't be bad for Intel.

And that's why the GigaOM piece was preposterous.

[Pishkado]

"There can't ever be a time when the world moves beyond x86?" They used to say that about horse-drawn vehicles. And you accuse Malik of "overwrought rhetoric?" Pot, meet kettle.

[pintos-n-cheese]

"There can't ever be a time when the world moves beyond x86?"

I concur with Pishkado. This one statement red flagged itslef as soon as I got to the period. What? Ntohing beyond x86? Talk about perposterous. Just about anything from the past 100 years will demonstrate that advancements will always be made in some shape or form, including technological ones beyond x86 archtiecture.

[Peter N. Glaskowsky]

Sure, eventually we'll have radically different computer systems, and they'll presumably have radically different ways to represent the programs they run. But we won't move to those systems just to get away from x86, and in any event, we aren't _already_ moving there as the GigaOM article implies.

[alxlr8]

Your ad-homonym attacks on other blog writers aside, I think your logic is flawed. On the one hand, you say that ISA doesn't matter any more, and that such arguements around the importance of ISAs were put to bed long ago. On the other, you say that "The value of instruction-set compatibility with the PC market will persuade developers of new cellphone platforms to go with x86 chips". I don't think you are right about X86 and its ability to penetrate this market, and I think the mantra that the internet runs on X86 is one of the biggest lies in the technology industry today.

As applications move onto the web, and more cloud based storage comes online, the level of abstraction at which these applications and services run increases. Additionally, the ability to view these services and interact with them on a mobile device is simply an issue of software that executes on that device. So it all comes down to how people will view the internet. In terms of devices, I think it will be up to ODMs and OEMs to decide which architecture is more relevant in the mobile sector, and at the moment there is one very clear winner.

Although I agree with you that anything that ARM partners can do to improve power, area etc. is also something Intel can do, there are certain aspects of an overall microprocessor architecture (not merely the ISA) that work in favour of ARM, irrespective of these process tricks. These certain architectural aspects feed into the micro architecture of an ARM CPU to make it more efficient for use at the mobile end of the scale. Atom will always suffer because - in acknowledgement to the efficiencies of a RISC design for mobile computing - Intel has had to attach a humungous CISC X86 decoder onto the front end in order to allow X86 to execute on the RISC core they have on the back end. This will always be required, and is something that becomes a mill stone around their efforts to penetrate this space with their ISA.

So architecture is important when fitting a solution to a market, and Intel will always carry around baggage that ARM will never have. Once the software story is sorted out, ARM will be a more compelling solution going into the future, and it is also why it is the de-facto solution in the present day.

Who are Freecell? Do you mean Freescale, the semi conductor foundry/SOC company that spun out of Motorola?

[mindgamez]

I am not sure what "ad homonym" means. Perhaps you should look up "ad hominem" before you use it in your argument. While I may not agree with the author of this piece I don't see any attacks directed at the writer of the GigaOM story, just at the stories themselves.

[pintos-n-cheese]

Speeliong mistakes are...mistakes! We all make them, so it's not helpful to point them out as if to attack a persons entire intellect, lncluding their recent argument, based, in part, on a spelling mistake. The fact is, you know what he meant!

[Peter N. Glaskowsky]

@alxlr8: Wow, that was a heck of a typo. I even slavishly copied the typo into the tags. Yikes. Fixed! Thanks.

As for the argument itself, the value of x86 compatibility with PCs lies with the compatibility, not the instruction set. I think that's clearly explained in the piece. If I'm wrong about x86's ability to penetrate the markets currently dominated by ARM, we'll see, but I'm not alone in the belief; at least three companies are developing new x86 cores to compete with existing ARM designs.

Your third paragraph is unclear. You refer to "aspects of an overall microprocessor architecture (not merely the ISA)" but then you offer, apparently as an example, the "humungous CISC X86 decoder", which is an ISA-specific issue, as if I hadn't just written about why that isn't a big issue any more.

Bottom line, the "baggage" of PC incompatibility is already killing ARM's competitiveness in netbooks, and this problem will start to give x86 an edge in smaller systems as time goes on.

I had several arguments I just didn't have time to address last night. Let me offer one now.

Apple sells a huge variety of iPhone applications through the App Store, all of which are designed to provide significant functionality through a relatively small 320x480-pixel window. Apple's Mac systems support Dashboard, where "widget" apps perform various functions in small windows. Both iPhones and Macs run the Mac OS X operating system, yet Macs can't run iPhone apps. There are several reasons why not, but instruction-set incompatibility is surely one of them. There's a good reason for that incompatibility today, but the reason will go away over time, and inevitably, either x86 or ARM is going to be pushed out of its niche. Which do you really think will win?

[alxlr8]

I fully understand that your arguement lies around compatibility, which is ultimately an architectural thing (and an instruction set thing). As such, the statement that ISAs aren't important is only true at the application level (hence abstractions possible in cloud applications), but it is not true at the hardware level.

In reference to your confusion regarding my statement about Atom and its decoder, the point you missed - and it is the main point of my statement - was that the old X86 architecture necessitates certain micro-architectural decisions to be made in order for it to fit the mobile processor socket being discussed. My statement to you is that those design decisions make any X86 based processor sub-optimal for that space. It is an architectural issue, it is an ISA issue, and it will be the very thing that puts people off wanting to use atom, if the choice was between atom and a more adept mobile processor architecture. Power is one of the biggest things mobile device manufacturers care about.

Intel has neutralised any silicon process advantage it had, by outsourcing manufacture to TSMC. Power will only get worse - not better - when intel moves to manufacturing atom through TSMC. Granted, they'll be able to produce SoCs, but their architectural disadvantages (in the baggage they must carry from X86 days) will make them a second-fiddle choice in the bill of materials race.

Ultimately, in software, PC compatibility becomes even less important as software moves out of the local processing realm and into the realm of distributed computing on the move. Does this mean that all applications will be cloud based? No, but it means the ones we use on the move most likely will.

[Peter Glaskowsky]

alxlr8, the point you missed is that those "certain micro-architectural decisions" become irrelevant years ago. Go read Weber's presentation. Read my post again. What I said was true. Stop with the hand-waving.

[man_w_balls]

There are still lots of PowerPC-architecture devices out there. I have at least 3 in my house alone (Nintendo Wii, Apple iBook, PowerMac G4) and some pretty powerful new ones are out there such as the Cell in the PS3 and the triple-core PPC in the XBox 360.
Will IBM, et all, stop production and all R & D into said PPC architecture? Doubtful. But I don't deny the usefulness and convenience of the speed of x86, considering I'm on a Core2 box typing this. From what I've seen though, Intel x86 kinda always sucked until the Core architecture series. PIII weakness, P4 heat issues, etc....

Generally, monopolies stifle growth and creativity. I say the more architectures, the better - some new innovation may be just around the corner, and the more people walking the block, the better the chances of finding it.

[Peter N. Glaskowsky]

But the instruction-set architecture is no longer the foundation of innovation, either. The x86 vendors (AMD, Intel, and VIA) innovate by adding new instruction-set extensions that are really nothing like x86. Heck, the AMD64 instructions functionally replace the basic x86 instruction set itself.

Innovation comes from microarchitecture and, increasingly, from system architecture. I have a list of dozens of changes I expect to see in system architecture in coming years. None of them depend on any particular instruction set. ISA convergence is a foregone conclusion at this point.

[sparrowhyperion]

I think it's rather narrow minded to think that the x86 architecture will be around forever. Back in the day, users and programmers felt the same way about the various 6502 based machines and their architecture, or the Motorola 68000 series CPUs, and their various iterations. For those who do not remember, the 6502 chip series were widely used in the older Atari 400/800 series of machine. The 68000 series was used in the later Atari ST series, The Commodore Amiga series and the first Mac systems. At the time, everyone was sure that these chips would remain dominant in the marketplace. Now however, the only use you might see these chips used in are some home appliances. The PC Clone as we know it today, became the standard because of it's attraction to businesses, and the generic nature of the instruction set. Programmers finally had a single set of instructions to write for which would theoretically work with all PC x86 based systems. However, as with the ISA, VLB, AGP, and soon PCI standards, it will eventually go the way of the dinosaurs.

I think the PC industry is due for a shakeup. Nowadays the x86 architecture is more a hindrance than a help. It's code is pretty much obsolete and in bad need of a makeover. as long as we keep a death grip on the x86 platform's instruction set, we will never be able to take full advantage of the full potential of current chip technology. Like the gas powered automobile, it has become a dinosaur which we are stuck with until someone has the stones to actually design and market something better. It may take years or decades, but I guarantee you, the x86 architecture will be replaced and it may not be nearly as long as one would think.

[pintos-n-cheese]

Agreed. Bravo.

[Peter N. Glaskowsky]

@sparrowhyperion: That's a good little speech, but you're confused about a lot of things. "Back in the day" there were dozens of different ISAs used in personal systems, and today, there's just one. I'm just saying the same trend will eventually sweep up smaller and smaller systems.

You even specifically say that instruction-set convergence helped to power the dominance of the IBM PC platform. Why would the x86 instruction set "go the way of the dinosaurs"? You just don't say, and that conclusion is directly opposed to your own supporting arguments.

If you'd care to explain why the x86 architecture is a "hindrance" or "obsolete" or a "dinosaur" I'd be happy to listen, but so far, you're only supporting my side of this discussion.

[kcotham]

@sparrowhyperion
Absolutely, the sooner we do away with the archaic monstrosity that is x86, the better.

[msjonker]

Wow dude, you need to read a history book. Things change.

[Peter N. Glaskowsky]

Some things change. Some things don't. Wheels are still round.

I don't think the x86 instruction set is good, never mind ideal, but it's become the low-friction way to write software.

[msjonker]

Wheels are still round because:
A. That is the definition of a wheel: "A wheel is a circular device that is capable of rotating on its axis, facilitating movement or transportation whilst supporting a load (mass), or performing labour in machines."
B. But to address the point I believe you were trying to make, wheels are still round because the laws of physics justifies them as being optimal.

Instruction sets and processors are a whole different story.

Now, I doubt that Netbooks are going to be the single changing factor, but that's no reason to say x86 will live forever.

[Peter Glaskowsky]

I'm just saying that x86 is, in effect, becoming part of the definition of personal computing devices. Just as it's not useful to make an out-of-round wheel (though people still try!), it's not useful in the long run to make computing devices with anything other than x86 cores.

[alxlr8]

I would disagree with the statement that "it's not useful in the long run to make computing devices with anything other than x86 cores". I would suggest that most computing devices on earth - from the ESP programmes in cars, to the Palm Pre - are all driven by everything BUT X86 cores. X86 cores are too noisy, or too power hungry, or simply too inefficient to ever be worth a look in such systems. Like it or lump it, the vast majority of hidden, discrete computing devices - including the way we compute in the home - will diverge away from X86, as the necessity to have compatibility with it becomes less and less relevant.

[Peter Glaskowsky]

@alxlr8: the reason most embedded systems don't use x86 is that they also don't use any other kind of 32-bit device. The average embedded processor core probably represents only a few cents of value in the product.

For these products, that quarter of a square millimeter of silicon for an x86 decoder is larger than the whole core.

As Moore's Law brings down these overhead costs of x86 compatibility, the use of x86 will expand into other markets that were previously not reachable.

Anyway, what I was saying about "not useful" was in the specific context of "computing devices", specifically "personal computing devices." And there is no possible way to say that a non-x86 architecture is an advantage in that market.

[nomadrider]

I don't really care about "full PC compatibility".
I just feel that anything portable, with a screen should have an optical drive in it.
That's my opinion of what "killed" the netbook. At least for me.
The platform isn't what's so important, it's what it can do that's important.

[Peter N. Glaskowsky]

That's a bit of a digression, but I don't mind. It's interesting you'd say that. I rarely use the optical drive in my MacBook Pro these days. When I get a new piece of optical media, the first thing I usually do is copy what I want from it onto my hard disk and pop the disc back out. More and more, I get data from other people, and from exhibitors at trade shows, on USB sticks or over the Internet. I can't even remember the last time I burned a disc for anything but backup purposes.

The one exception here is DVD and Blu-ray movies, which are still best distributed on optical media, and even though it might be more convenient to copy them to disk for viewing, that generally isn't authorized by the licensing terms. (Something that ought to change, if you ask me.) So an optical drive is still useful for watching movies. I don't do that much, but some people do, and I suppose they, like you, would prefer a small notebook with an optical drive. But I don't think that "any" portable computer should have one. For many people, the size and fragility of an optical drive are problems that outweigh the benefits.

[kcotham]

x86 is an archaic architecture that is technically inferior to RISC architectures. The only reasons it has done as well as it has is: it got a huge head start in the numbers game (so many machines had one in it) and they've been able to compensate for architecture shortfalls with raw clock speed.

PowerPC, ARM, and the like are more elegant designs and will eventually take their place in embedded and mobile processing respectively.

[Peter Glaskowsky]

That was a great argument 20 years ago, but it was a weak argument 10 years ago, and fully obsolete 5 years ago. Sorry.

[alxlr8]

ARM already ships in 25% of the digital devices sold in the world., so I don't think there is any element of "eventually" about it - they are already there!

[mycall]

x86 has already moved beyond x86.

We have x86 but we also have MMX, SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, SpeedStep (EIST), Intel 64/x64-64, XD bit, TXT, VT, Hyper-Threading, Turbo Boost, QPI, Smart Cache, and HD Boost (just to name a few things)

[Peter Glaskowsky]

Okay, I agree with that. But it still means that only x86 can replace x86. :-)

[emfle]

One reason x86 is probably doomed for small systems is that the margins there are so thin and the volumes so high, that even the slightest technical disadvantage can cost many millions of dollars in profit. The compatibility may make up for it, but do you really expect people to run unmodified Windows applications on these devices? (And why doesn't backwards compatibility with iPhone apps matter?)

Also, if those low-power "tricks" are so easy to apply to the Atom, why haven't they? The Atom still uses many times more power than high-end ARMs.

I suspect Atom is probably a controversial product within Intel. They would much rather sell Nehalems instead. Low-margin products like the Atom is not a good match for expensive fabs, so now they are outsourcing production of them to TSMC. But if they are not using their fab advantage, there is only backwards compatibility to fall back on, and that may not be enough.

[Peter Glaskowsky]

But the whole purpose of the Intel-TSMC deal is to ease the margin pressure on x86-based systems to the same level as ARM systems. Yes, I do expect people to run unmodified Windows applications on these devices. They do that now, on MIDs from OQO and Fujitsu and other companies; see the MID article I linked above. Over time, there's definitely value to cross-platform binary compatibility right down into the cellphone market.

[ctfoley]

People don't want full pc compatibility on a netbook. What, are they gonna run Photoshop and AutoCAD on it? all they need is the web. oh, and arm has a pretty good head start on intel, wouldn't you say?

[Peter Glaskowsky]

Well, no, I wouldn't. Netbooks mostly run Windows these days. Linux netbooks mostly get returned or upgraded to Windows. People DO want PC compatibility, because that means a lot more than Photoshop and AutoCAD. Obviously.

[ctfoley]

Yes, there are a lot of lighter-weight Windows programs you could run, and once you get into OneNote/PowerPoint territory, web apps are still inferior. But are they inferior enough to give up a lot of battery life? for some people they are, but for many they aren't.

[alxlr8]

I think in this rather young market, Intel will start to see the power of ARMs licensing strategy - and the flotilla of companies who brandish it - with pretty good penetration into the computing markets, including netbooks. I wouldn't bank on customers going into the local computer store and really caring whether or not their netbook is windows compatible in two or three years time, because by then I suspect ARM processors will be running both Linux and Windows, then it will be a choice based on battery life and innovation.

[Peter Glaskowsky]

Well, Intel has its own licensing strategy now, and I have yet to see an argument here as to why that doesn't drastically undercut ARM's ability to sell into smaller devices. But there are sure some strong arguments why people don't want ARM processors in netbooks.

[3rdalbum]

The technology media like to write two things

1. Cloud computing will take over
2. Wintel will always be the dominant platform

If cloud computing does take over, it doesn't matter whether you're on Windows on x86, or Linux on ARM, does it?

[Peter Glaskowsky]

Let's just say you haven't seen me saying much good about cloud computing. It's good for some applications, but I see no possible way it can "take over" much of the overall computing market.

[kkrewell]

Peter,

I have to disagree. In order for Intel and x86 to be successful in the smartphone market, there needs to be a reasonable need to run the PC operating system and programs on a 3 inch screen. Otherwise, x86 compatibility is meaningless in a market where ARM-based mobile operating systems already dominate. The x86 instruction set and traditional Windows has no inherent advantage in the smartphone market. Apple has shown that a mobile operating system, even one based on a desktop core, needs a different user interface and application design.

To Peter?s point that instruction sets are not relevant, why must x86 win in the end? There are only three x86 vendors and only one of those is both financially healthy and shipping significant volume. While there are a few hundred million x86 processor shipped every year, there are many hundreds of millions of ARM-based designs in cell phone, microcontrollers, and the like shipped every year (over 10 billion since 1990) from many different companies, large and small, and in between. ARM is not some obscure processor, and it is used by many innovative companies (according to ARM, 60 partner companies were at MWC'09). For Intel (or AMD for that matter) to take x86 into cell phones is proof that if your only (successful) tool is a hammer, every problem looks like a nail.

Intel?s agreement with TSMC to offer an Atom-based (Intel?s lowest end, lowest power x86 core) hard core for SoC design is Intel?s recognition that it needs to tap into the ARM ecosystem. The same ecosystem it failed to win over with XScale. Details of the agreement are still meager, but I?m concerned because it allows Intel to vet who and what will use the Atom core. That is not open enough for true innovation and competition.

Of course, these are rational arguments and the market is not always rational. If it was, x86 would not have beaten out 68K and Z8000 in the 16-bit instruction set wars.

[Peter N. Glaskowsky]

There's no "need" to run a PC OS on a small screen, though as you know, Mac OS X _is_ the foundation of the iPhone operating system. In fact, there's no "need" at all; there's merely value.

It's valuable to run x86 code on handhelds because so many little bits and pieces of the code can be shared between handhelds and PCs: multimedia codecs, browser plugins, desktop widgets, development tools, etc.

This kind of commonality also reduces the total effort needed for debugging, interoperability testing, etc., which creates another kind of value for x86.

When these values exceed the costs of x86 compatibility-- which they did ten years ago in the PC, workstation, server, and laptop markets-- companies put their efforts behind x86. We all knew that the x86 architecture wasn't as good in most technical respects as the 68K architecture, but it was obvious from the day IBM announced the first PC that business issues were more important than technical issues.

I don't see how the Intel-TSMC deal represents an attempt to "tap into the ARM ecosystem". It's an attempt to compete with it, and it's likely to be effective.

I don't think your arguments are rational, but I think it will be entirely rational for many companies to adopt the x86 architecture instead of the ARM architecture for smaller and smaller products over time.

[sk00n]

Lol, man where did You live for last 5 years? I have been bound to the companies dealing with arm, building solutions on arm, conquering the world with arm. What wintel means to somebody who has been bound to arm for few years? Bluescreens on humungus ad screens in big capitals wasting the space and cash of the customers of ad agencies, bluescreens in the atms, bluescreens on the tubes in all known capitals and different public transport means throughout the europe, the microarchitecture so severly constrained they cant move beyond the power consumption of atom (didnt they try? I am afraid they did but with 200 pound of baggage is not easy to be agile), negligible market share in smartphones, portable devices, mp3 players, mobiles, automotive, servers (linux not windows), shall I say more?

As to the macos You might not realise that: 1 this is based on linux, 2 the main power behind it is the ui + mechnical design rather than os itself.

As to the software used in automotive if You mean the onboard computer for ignition and fuel injections, You are mostly right, for anything on top from media player, through radio, lcd driver, speakers and multimedia systems+gps You are plainly wrong. Yes they do run on 32bit archs, no its not intel or x86.

And if You knew more so much more about the assembly and possible instruction decoder for both, You wouldnt say 2 square milimiters is the problem (its much bigger area of the wafer definitely).

And few more questions from a geek: why didnt wintel succeed in smartphones and mobiles in general (hint: power issues consumption caused by wierd/crappy/suboptimal micro arch)? why not so successful but important project as xo is running awawy from intel (hint: power issues^)? why theres no important netbooks running on windows (hint: the startbutton takes 0.25 of the screen + power issues)? Whats the volumes of intel sales in personal use devices and whats arms?

And misreading amd presentation is a bad sign of Your general journalist work, what they meant is arch is not important, meaning people will be more and more eager to switch to other things + cloud apps are neither constrained to single arch Which gives a huge yes! to the other archs esp arm.

Please try to realise that intel is trying to defend its base (notebooks+smartphones+servers) with pocketnife while being raided by the hordes of ambidexterous robotic lightsabre operators rather than trying to conquer the market reserved to arm and mostly linux.

I will become an avid reader of Your writings, I suppose.