they're still tied to Windows which is increasingly weaker and more rotted over time, losing market share every year for a decade (Microsoft doesn't even seem to care afaict)
the Apple M1 laptop is an incredible value, offers 10x the performance the average person will need, with good battery life, a better OS/ecosystem...and its priced right!
Intel needs to seriously look at Linux...once Windows enters into a death spiral in a few years (seems inevitable now and Microsoft seems to not care as long as Azure keeps winning), what else will they have?
start now with serious price cuts....I don't need to replce my Linux laptops, but would for the right price
Cheapest one you can currently buy is the Air. That starts at 1k USD pretax. That's not within reach for a large majority of the worlds population. Not even within reach for a lot people in richer western countries.
Most people need a Webbrowser that can do Email, Streaming, and some social media.
What makes you think windows is in a death spiral? They still have a good 75% market share and while that fluctuates, it doesn't seem to really drastically decrease.
> What makes you think windows is in a death spiral? They still have a good 75% market share and while that fluctuates, it doesn't seem to really drastically decrease.
Of what? Most of what I see normal consumers doing is in Android, typically on ARM. I think my observations are a little biased, but not a lot. According to statcounter:
Android 41.56%
Windows 28.29%
iOS 17.29%
OS X 7%
Unknown 3.05%
Chrome OS 1.32%
Critically, it's no longer a monopoly; there's a lot less lock-in than there once was. Even Microsoft Office is gradually moving to the PWA model. Most servers and embedded systems I see run some variant of Linux as well. The market Microsoft has 75% share in seems to increasingly by a niche market within a much larger ecosystem. Microsoft seems to have pivoted, and as someone pointed out, no longer cares too much.
However, the Wintel monopoly is dead. If Intel is tied to Windows, they'll increasingly be swallowed alive. TSMC manufactures most of those ARM chips. There's no way a smaller player can keep up in fabs and manufacturing technology.
Intel is in the CPU market, not in the "corporate desktop" market.
If you'd like to see how well trying to focus on that niche works, have a look at DEC Alpha, MIPS, PA-RISC, Itanium, Sparc, PowerPC, etc.
At the end of the day, it comes down to manufacturing volume. Volume leads to investment in next-generation technologies. More investment means, at some point, competitors can't keep up. Manufacturing doesn't care too much about who the ultimate end user is.
His point was trying to measure OS market-share by checking desktop-only is no longer a valid approach.
Even big players like Microsoft are forced to offer their software and services on non-Windows platforms (like mobile and tablets!) if they want to remain relevant.
So counting desktop only paints a false picture, wrt where users are and where computing is happening.
And that last bit is what is relevant from a Intel POV. An increasing amount of compute is now done outside the Intel-chip arena of desktops, like it used to be before.
These industries won't buy macs as they are too expensive to maintain and to integrate into their existing active directory domains. (With a few exceptions for jobs where the Mac is the optimal device like graphic design)
No idea about percentage but windows won't disappear soon from those.
The android market share is mainly mobile devices and Microsoft gave that market up a few years ago, focusing on azure and office.
Here again, focusing the corporate/gouv world.
The corporate/gouv world has many windows only app and processes that would cost a fortune to replace.
(Not sure if your last sentence was a joke meant to undermine your argument preceding it? Maybe by mainframe though you're referring to servers.)
I generally agree with your point though but I am seeing that as more and more of office work moves to living within a web browser, it is becoming more hardware agnostic.
Where I work has applications dating back decades with complex data analysis algorithms developed by very smart people. Strict backwards-compatibility is critical. We have mainframes.
There are many other reasons to run mainframes. For example, mainframes offer pretty incredible features with regards to redundancy, and making everything hot-swappable in case of failure without interrupting workloads. That's especially useful in places like the financial industry. There, a single failure or error can cost much more than a mainframe.
They're a niche, but a big enough one to support a multi-billion dollar business. It shrunk for many years, but it's currently even growing.
I was doing work on an IBM AS/400 mainframe in my previous job, just over a decade ago. It supported Java and Tomcat (which was fairly current at the time) out of the box. The only oddity I needed to deal with was making sure I had a proper JDBC driver for its DB/2 database.
Don't get me wrong, I'm glad I don't have to deal with it anymore, but mainframes are definitely still around.
Back in the day, they cost a fortune to set up and replacing them is a high risk - low value for many CEO's.
In Québec, the SAAQ (provincial DIV) decided to get rid of the mainframe by replacing it with SAP and with the amount of data they had, the migration took them a month of downtime.
I'm sure that Windows Server will gradually be replaced by containerized app running in some cloud but windows desktop will take a long time to be replaced, especially because of office.
"These industries won't buy macs as they are too expensive to maintain and to integrate into their existing active directory domains. (With a few exceptions for jobs where the Mac is the optimal device like graphic design)"
I'm not convinced. About ten years ago one could say exactly the same thing about Macs companies located outside of California. I only had my first work Mac in 2012, when I convinced my boss to let inherit one from a graphic designer. Since then, I've worked in two extremely traditional corporates that gave me the choice to use macOS. And plenty of executives were/are using macOS too.
In the meantime things got actually easier than 10 years ago: web apps became prevalent and even Microsoft joined the party.
You can buy an M1 8/256 Air at $800 right now. I'd expect to get 5 years out of it minimum (we're buying one shortly to replace a mid-2013 Air). That's less than $15/month if it gets thrown in a bin in 5 years. It's under $20/mo if we bin it in 3.5 years.
I feel like the Chromebook on the lower end is great for a lot of uses/people and the Air is an easily achievable next entry-level step up.
Looking at the best selling laptops at Microcenter, they're $600, $700, $700, $1100, $800, $800, $750, $900... Having an M1 Air slotted in at $800 seems right in-line with the market.
Especially if you plan to use it for next 5 years. Perhaps if you're sure you don't want to do anything other than web browsing, but then maybe you don't actually need the M1.
BTW, the cheapest M1 Air (8 GB RAM) in Austria is 1200 EUR.
I tried to do web dev on an 8GB Mac and it wasn't fun. If you use an IDE, it will be constrained for memory if you have things like browser windows open, you will have a bad time if you try to open multiple projects and running Docker made it all worse. Virtual machines were out of the question which made mobile app development difficult.
If you can stick with a terminal or some lightweight editors, 8GB might be okay for you, but I feel like 8GB has been pushing it for at least 5 or more years now.
There was a lot of discussion of this at the time the M1s first came out, with loads of people (who had never touched one) confidently declaring that 8GB of RAM was far too little...and then nearly everyone who actually bought an 8GB M1 Mac saying "yeah, it might not handle heavy loads well, but that's what the Pro machines are for—in normal use, it has no problem at all with 8GB."
And then a bunch more people talking in a fair amount of detail (that kind of went over my head) about what Apple did to improve memory management with the M1 so that this would be possible.
So don't assume that 8GB on an Intel Mac is the same as 8GB on an Apple Silicon Mac.
Now, it does sound like you were trying to run a whole bunch of rather heavy stuff on it—y'know, the stuff that one might call "pro", and possibly even buy a "pro" machine for? The kind of thing where if you're buying the absolute bottom of the line, you might be able to expect that it wouldn't work well.
I don’t assume, I currently have a MacBook Pro M1 with 8gb of ram.
It all depends on what kind of tools and IDE you have. When I avoid a having 50 tabs on chrome and only the apps I need it’s perfectly usable for web dev. Of course 16go is more confortable but I’m working with 8 with no problem on light web dev.
I honestly think it was all just hype, or some sort of FOMO of every reviewer not jumping on the M1 hype-train.
I own a base Macbook Pro 14", 16GB RAM. If anything, it feels like less than 16GB (maybe because the shared video memory or macos bloat) saved by the fast SSD that's inevitably going to fail not too far and turn the laptop into e-waste.
I feel like every single talk point of the early review cycle was a huge exaggeration:
1. battery life just browsing the web is no more than 4/5 hours. Like any other windows laptop.
2. Performance is really nothing special really. Things like maximizing/minimizing a youtube video are choppy and lag for 2/3 seconds.
I mean, it's still the best value in the market for the screen, keyboard, chassis, speakers combo but Apple Silicon is extremely overrated.
> Now, it does sound like you were trying to run a whole bunch of rather heavy stuff on it—y'know, the stuff that one might call "pro", and possibly even buy a "pro" machine for? The kind of thing where if you're buying the absolute bottom of the line, you might be able to expect that it wouldn't work well.
That's nice, but I can buy a much cheaper non-"pro" computer with better specs that can handle a basic workload that involves a browser, IDE and maybe Docker. Like I've run all of that on entry level machines without a problem.
Because Apple will forcibly kill it after that with incompatible software updates. Windows machines keep going until they physically break.
But what everyone here is missing is that the consumer desktop/laptop market is dying itself. Soon it will only be for enterprise uses and gaming. Most people do all their computing through a phone or tablet. I know people who don't own a laptop.
> Windows machines keep going until they physically break.
That's interesting, as I just this month replaced the two kids' Dell 9020s with new PCs because the 9020s didn't support TPM 2.0, required by Windows 11.
"Most PCs that have shipped in the last 5 years are capable of running Trusted Platform Module version 2.0 (TPM 2.0). TPM 2.0 is required to run Windows 11, as an important building block for security-related features. TPM 2.0 is used in Windows 11 for a number of features, including Windows Hello for identity protection and BitLocker for data protection."
The 9 year old Dells were running just fine on Windows 10 (same as the 9 year old Air was on Catalina), but upgrading to the latest OS in both cases required a purchase.
I'm curious why you felt you needed to upgrade. To my knowledge Windows 11 doesn't do anything that Windows 10 cannot, except for support Intel e cores.
They wanted to play Valorant with their friends; the anti-cheat code did not work on the 9020 on Windows 10; there was a limited amount of time I was willing to spend to find a cheaper workaround for a machine that’s 9 years old, with the prospect of doing the same again 6 months later.
That was the actual driver, though this thread is about OS vendors forcing changes in order to run the latest OS, which is why I focused on the Win11 TPM2 requirements and mentioned the 9 year old Air running a still-patched Catalina.
I don't believe you needed to upgrade to windows 11 for valorant compatibility. You just needed to be on a recent build of windows 10, downloaded from windows update.
Maybe. We always have Windows Update turned on, so it was almost surely up-to-date and I spent a period of time over several different evenings trying to get this to work. Whatever error message came up when jumping into the gameplay part of the game led to the TPM issue being the root cause. Ultimately, we'd gotten our 4 years of value out of the Dells (4-ish years old when we bought them) and decided it was reasonable to replace them.
To the point of this sub-thread, it's still an OS vendor forcing a purchase to run the latest OS. 9 year old Macbook running Catalina and 9 year old Dell running Win10, both unable to update to the latest OS seems very close the same to me.
All my Windows machines from 2009 are still pretty much ok, while having been upgraded to Windows 10, came with Windows 7, and went through 8 and 8.1 as well.
> Because Apple will forcibly kill it after that with incompatible software updates. Windows machines keep going until they physically break.
This is demonstrably false. My wife continues to use my hand-me-down MacBook Air I bought in 2011. It was supported officially by the macOS until the most recent release, so she’s only one OS revision behind. I also recently retired a 2009 Mac Mini that I used 24/7 as a home server, only because the power supply failed. I used the G4 Mac I bought in 1999 as my daily driver (with some upgrades) until 2009.
A decade or more of use is an incredible value, and I’ve done that multiple times with my Macs. Few of my PCs have had a useful life approaching that length of time (the closest was an Atom machine I ran for 7 years).
Same here. I still have the same, 2012 11" Air, and it's my "travel" laptop. It had a battery replacement a few years ago (and Apple did for free!), but other than being a Netflix machine, I still code low-level stuff and frontend in it.
It was in 2018, so after only six years. Five years ago, time flies... And the battery was swollen up, it actually damaged the trackpad. I went to the store to get an estimate, but they said they'd do it for free.
I wonder if it was a recall. I didn't speak German that well back then so it might have gotten lost in translation. I recently also had issues with my Airpods Pro and they replaced for free, but they told me it was a recall.
I also had a swollen battery in a 2013 Macbook Pro that was eligible for a free replacement in 2018 or 2019 (I forget exactly when, but it was before the 2019 16" Macbooks came out).
It was covered outside the normal warranty because of the swelling issue. Mine swelled to the point where the trackpad was slightly unreliable.
I'm not sure "forcibly kill" is the adjective/verb I would have used. I have plenty of older Macs still running older versions of Mac OS. In some ways I kind of prefer them (because I still like the older Mac OSes).
Maybe this is more of an indictment rather than a defense of Apple, but one machine I keep on an OS that's over 5 years old because it has the one version of Apple's Photos that didn't yet screw up the retouch tool. (And I use that tool exhaustively in retouching old family photos I scan in.)
My MacBook Pro was supported by OS updates for a DECADE. I have a MS SURFACE Laptop Gen 1 from 2018 I can’t update to Win 11 — and will EOL in 2026. That’s from the flagship microdot branded laptop.
I have 8gig and its completely fine for webdev work. Its faster machine than my desktop. It has to be completely fine for majority of work people do on computers.
You could buy a laptop with 16GB to 32GB of memory and more than double the storage space for $800, and you can buy laptops that you can actually upgrade the hardware on. You can also spend less and get a similarly spec'd laptop. I just picked up a new laptop with 16GB/1TB and 8 cores for $500 and upgraded it to 64GB of memory for $100.
I bought a low end Mac in 2015 with the same 8GB/256GB specs, 8GB wasn't enough even for browsing/chat/media/etc at the same time, and the storage space gets eaten up by the macOS base system and apps.
I have an 8GB/256GB 2015 MacBook Air. I use it for web browsing, messaging, spreadsheets, watching media, email, and photos. It has worked fine for me the whole time, and I usually have it connected to an external display.
I doubt 80%, maybe even 90%, of the population needs anything more than that.
If all that matter to you is RAM and HD space, you can as well buy a cheap desktop tower. As for "similarly spec'd" laptops if you put good perf and low consumption in the equation, they just do not exist. Yes Intel Mac were not extremely interesting at least on the CPU + GPU + RAM + HD + power side of the equation. The M1/M2 is - esp if you also value a decent screen. And the M1 + 8GB is reasonable depending what you are doing with it.
> That's less than $15/month if it gets thrown in a bin in 5 years
My mom uses a cheap Dell laptop that is 10 years old. I bought at the time for around 350 dollars is memory serves (it was in a different currency, so it is hard to do the conversion).
She is non-technical, and it absolutely suits her needs to this day. Only thing I did in terms of maintenance was buying a new battery.
The cheapest Mac just won't get the same cost/lifespan ratio.
Keep in mind that the vast majority of users are non-technical.
> You can buy an M1 8/256 Air at $800 right now. I'd expect to get 5 years out of it minimum
Some people simply do not care about tech to a point where almmost any price is too high. I just had to persuade a relative to spend $250 to replace a 16+ years old Intel Core Duo with an HDD. It takes 2 minutes to open a browser window on that thing and she didn't even consider that a problem. The replacement will probably last until it refuses to boot.
Macs are incredibly good value. The M1 Air actually starts at 800$ if you chase the deals and $900 if you buy from Apple with education discount.
Nothing ever comes to the quality at that price, so even if you can buy much cheaper laptops those suffer in quality and performance. This means the value per dollar is not any better than MacBook Air, unless you are extremely short on dollars.
Also the re-sell value of Apple products is good and combined with the durability of Apple laptops(this is where the repairability folks fail to convince most people because for most people Apple devices don't break down, they become obsolete after extensive use) the money spent on Macs turns to be an investment. The "Apple premium" you pay when you buy an iDevice is actually a recoverable when you sell it.
I agree they are but what if you can't afford to spend 800 or 900 on a laptop?
There's also plenty of people who use their laptop until it's dead. Infact, I'd argue it's probably more economical to purchase a repairable laptop and replace parts as needed.
If you can't afford $900 you shouldn't be buying a laptop unless you absolutely need it in short term to get out of your current situation. You are probably in a precarious state, not really a chooser at this time and all you can do is to sort by price and buy the cheapest.
Thats an incredibly privileged view on this. The world doesn't run high paying western salaries. Most people earn significantly less. The average monthly indian salary, by now the largest country in the world population wise, is roundabout 450 USD.
Your advice boils down to just get richer and don't be poor.
My advice is, don't spend your money on low quality stuff unless you absolutely have to(i.e. maybe you will do some data entry stuff to buy food, then maybe you can spent a few hundred dollars on a laptop but if you are not in that kind of a situation save up buy something of good value later).
Not massively advertising low cost options and not having any at all are wildly different things.
Of course companies push the items with the highest margins. They want to make money after all, but they still have plenty of options in the sub 600 category that will last 5+ years under normal use without issue. There's obviously a market for low priced laptops, otherwise so many companies wouldn't build them. And it doesn't hurt Intel. If the choice for them is to sell a low margin CPU or none at all, they're gonna sell the low margin CPU.
AMD's 5nm chips compete with and outperform the M1, low end laptops ship with 16GB of media by default, 8-10h of battery life is good enough for most people and Dell XPS laptops, for example, weigh 1.17kg, which is less than the M1 Air.
That's significantly slower machine than M1 Air. The screen is much worse, the actual battery life appears to be half of the M1 Air and there are all kind of other things like the sound quality or the trackpad that fall pretty short.
I'm glad you are happy with it, for 500 bucks seems like a fair deal, but this is nothing like Macbook Air.
The M1 MacBook Air lasts over 12 hours on battery and runs silently and cool while being fanless. I purchased mine new on sale for $750. Where is this magical laptop that beats the M1 Air?
Thing about ThinkPads is that there's almost always a premium option for hardware like screens that can be dropped in with nearly tool-less effort. I have an old X250 that I recently upgraded from the crappy 1366x768 shipping display to a new 1920x1080 for $60. The real problem with that model (besides the ancient 2 core CPU) is there's only a single RAM slot, and anythinng more that 8 GB is hard to find (and way more expensive than warranted).
I run an old (I think maybe 4 or 5 years) thinkpad with 64GB RAM. I see often these claims of superior performance of new macs, and it leaves me scratching my head. I do not have (almost) any workflows where the performance is not sufficient. For the ones where performance is not sufficient, I am happy to turn on my desktop. So my question is, what are the workflows people have benefit from the superior performance of new macs and at the same time do not justify working on a beefy desktop?
> What makes you think windows is in a death spiral? They still have a good 75% market share
The biggest threat to Windows isn’t Apple and Linux, it’s mobile phones. People aren’t buying computers like they used to. Most day to day things can be done on a smartphone now.
I see this first hand in my own family. My wife and mom both use their phones for pretty much everything. Email, web browsing, etc. If my wife needs a bigger screen for something like a yoga video, she usually uses a TV in the house which has a Roku, chrome cast, or similar attached.
> The M1 is a good chip, but Macs are expensive. Cheapest one you can currently buy is the Air. That starts at 1k USD pretax.
Get the MacMini. The low end M2 Mac mini is $599.00 USD right now, although the one I run for my desktop is the more expensive one at $2,199 with 32GB, more cores, more gpu and more drive space.
Price cutting by enough to pull your demand forward is probably a losing strategy for Intel. If they cut their CPU price in half, would they sell twice as many? I doubt it.
Their gross margin is between 40 and 50%. Call it 50% to be generous and make the math easy. If their selling price today is 6X and they sell U units, their gross profit is 3 * X * U. If they cut the selling price to 4X and sell 2U units, the resulting profit is around 2 * X * U. If they were somehow able to move 3 times as many units with a 33% price reduction, their gross profits would stay about the same as today.
What would a 33% reduction in the price of the Intel components in your laptop mean at retail? A 10% reduction in the shelf price? 15% maximum?
The x86 ecosystem is far superior to Apple Silicon, and that is not going away any time soon.
Macs have one big advantage - efficiency, which translates to long battery life and more importantly for me personally - NOT OVERHEATING.
However, porting software to apple silicon results in inefficiencies, and for developers it gets a bit trickier.
I have a M1 Max with 65GB ram and I love it, but the moment Intel comes out with chip that doesn't overheat and is as performant, I'll be buying an Intel based laptop (granted, I will have to wait for longer).
In a perfect world, Apple would get back to offering Intel as a CPU option when Intel comes back to the leading edge. You'd get the Mac OS, x86 and performance in a well built laptop - a dream!
I was referring to personal computer usage. Yes, usage specific chips can be a better option, but as far as general processing/usage is concerned, I'd take x86 over Apple Silicon any day of the week.
Overheating is entirely down to the profiles the manufacturers use. Intel processors can be efficient, but usually they get clocked to the moon on desktop. Laptops are so all over the place it's impossible to judge them either :/
I haven't tried limiting clock speeds to keep the laptop cool, and it seems a bit crazy investment to buy a laptop to try that, but I do have an old MBP with Intel that I can test this with.
You won't see power efficiency and temps comparable to apple silicon on the Intel MBPs, being 14nm, though you can probably get a nice marginal improvement like giving up 10% performance for 20% less power draw. The newer chips from AMD and Intel are comparably efficient to apple silicon with a restrictive power envelope.
Does it have to be Intel, or would you consider AMD?
Yes, if you look back even just to Zen / Zen+ laptop chips, they were not impressive, but Zen 2 and newer based chips (Ryzen 4000 or higher) are performant, efficient, and do not overheat. There are some crazy 7045HX chips that are basically desktop chips that you wouldn't want in your laptop, but the 7040HS line is exactly what you were waiting for (as well as the 6000U and 6000HS line). Unless you need Intel over AMD for some reason?
>losing market share every year for a decade (Microsoft doesn't even seem to care afaict)
Of course they care.
Windows, at one point, had 95% market share. Do you believe that to be sustainable? Do we want that? No, in both cases. They are still dominant in desktop computing.
>offers 10x the performance the average person will need
Why is buying something 10x more powerful than you'd need at 2x the price "good value"?
>a better OS/ecosystem
This is debatable. I don't find MacOS to be all that great, personally.
>once Windows enters into a death spiral in a few years
You should have a look at Windows' desktop market share and get back to us. They still own the space.
I commented on thus somewhere else, but it seems fitting here as well: my non-technical parents love the tabloid news that Windows shows, and they love that candy crush is pre-installed.
I suspect most users are not the same as the HM crowd.
Most mobile phone users aren't in the HN crowd, yet the iPhone continues to sell well. Why is this true, when a $199 unlocked Android is more than sufficient?
> I would argue in the same way that buying an expensive pair of boots that last longer than a cheap pair to be good value.
While a fair argument, I think its also fair to counter that the conditions in which you may use your boots don't arbitrarily change and get less efficient and more difficult over time. Sure maybe sometimes you'll be hiking through mud but that will dry out and get back to normal, whereas software might turn into quicksand as hardware gets older. Future proofing your hardware is effective, but only to a point.
> The reason that the rich were so rich, Vimes reasoned, was because they managed to spend less money. Take boots, for example. He earned thirty-eight dollars a month plus allowances. A really good pair of leather boots cost fifty dollars. But an affordable pair of boots, which were sort of OK for a season or two and then leaked like hell when the cardboard gave out, cost about ten dollars. Those were the kind of boots Vimes always bought, and wore until the soles were so thin that he could tell where he was in Ankh-Morpork on a foggy night by the feel of the cobbles. But the thing was that good boots lasted for years and years. A man who could afford fifty dollars had a pair of boots that'd still be keeping his feet dry in ten years' time, while a poor man who could only afford cheap boots would have spent a hundred dollars on boots in the same time and would still have wet feet. This was the Captain Samuel Vimes "Boots" theory of socioeconomic unfairness.
If they care they sure don't act like it. Microsoft is clearly squeezing Windows users for whatever revenue is left. Nothing else explains the user hostility, backwards compatibility breaks and unending advertising and data collection.
I always wonder how people meet enough random samplings of young people to get this information!
I have 9 nieces and nephews, plus some neighbor friends that have kids. Of them, maybe 2 or 3 use iPads. The ones that have laptops (other than school provided) have them for PC gaming, and they run Windows. I guess I'm in a different area of the U.S. from you, based on my anecdotal data.
My parents are not technical but view pretty much all changes since Windows 10 as downgrades and annoyances. Things in the UI change randomly for no reason (e.g. copy/cut/paste are now unlabelled buttons placed in an unexpected spot of the context menu), the dark theme is buggy and can make the minimize/close buttons invisible and the few areas not touched by Windows updates were absolutely ruined by Dell. You may think the latter isn't connected to Windows, but you can't buy a Dell with Mac OS.
This is the real consequence of Intel being run by MBA sociopaths for two decades now.
Intel (and AMD) needs a good consumer OS to keep it relevant in the next decades. It's what keeps the DEVELOPERS DEVELOPERS DEVELOPERS on x86, and therefore use x86 server architectures. Either that, or it is a commodity ARM business.
Microsoft has never been up to the task of a good consumer OS. Intel always should have had multiple irons in the fire, and with Linux they had a built-in weaponized competing consumer OS. They delegated all leadership to Microsoft rather than keep irons in the fire. WHich is ... ok ... when MS is the big monopoly on the block. That doesn't exist anymore.
If Apple pre-iPhone produced OSX out of probably less than 100 or so engineers... Intel could have done that with 1/1,000th of their quarterly revenue. Or less! It's just spitshining an existing, working kernel, and adding a bunch of testing for the cacaphonous PC hardware/peripheral market that they were already at the nexus of.
Apple had revenues of 6 billion in 2001, the first year of OSX. Intel make 26.5 billion in "a bleak financial year".
It's not too late! It's still peanuts to do a spitpolish of Linux and do hardware testing/driver development for desktop linux! AMD or Intel (or a shared foundation) could do it! It could probably happen within a month. Partner or acquire ubuntu, or Suse, or ... anyone.
This isn't just insurance. Intel has had DECADES of frustration getting Windows to fully use/support its hardware and CPU features. A lot of the time Windows would hold people hostage by waiting a major OS release (like for USB) before supporting something.
Not with Linux. Boom, your stuff is supported as soon as you integrate it into your distro.
This is the path to a key thing that Intel doesn't realize it lost: relevance. Microsoft has decided Windows isn't really relevant. That was Intel's only path to relevance.
All the things they wanted to stay relevant in: x86 on phone, SSDs in PCs, Optane, etc. The path to that relevance was having a usable consumer OS. That you had good control over. You get your hardware demo'd and supported at maximum speed.
Don't like Linux? Well fine, pick up OpenBSD, or whatever remnant of BeOS is out there. RedoxOS? whatever.
Gelsinger probably is following a path of "get engineering" fixed. Well, ok, good strategy given how bad the MBAs fucked up everything. But Intel isn't the big daddy and that doesn't really solve the problems, it just mitigates the biggest bleeding.
Intel needs SOFTWARE RELEVANCE. Intel does not have that anymore. Intel still has the resources and marketing power to achieve that. THE WINDOW IS CLOSING FAST. Regain software relevance.
Agreed, but I think AMD can pivot to an ARM architecture. Actually, Intel and AMD would be idiots to not have an Apple M1 chip competitors in active development.
Intel has a lot more denialism and inertia in their management to fully embrace / not undermine a legit ARM product.
This is largely why they hired Alyssa Fitzpatrick from Microsoft, and where the "Intel Accelerated[1]" software partnerships program came from. I don't know if it will work or not, but it's a start.
>Why is buying something 10x more powerful than you'd need at 2x the price "good value"?
M1 performance won't be more than you need for long. When the dev community finishes transitioning to M1/2, that will be the performance baseline for lazy websites and electron apps. Intel laptops will have their CPUs pegged to 100% constantly when running more than one tab in Chrome. We're already seeing this at my company. Intel performance regressions are discovered the last minute before release, because everyone is using M1. And the solution is usually just to turn off the offending feature for Intel Macs. It's legacy.
I think normal people (ie non-developers) are less concerned about the performance, as the battery life and the need to almost never need a fan (and keep it on your lap comfortably).
You can commonly find Macbook Airs at $800, and I doubt many are buying sub-$400 laptops, unless they are going with low-end Chromebooks, in which the Chromebook vs Windows is a reverse spin on whether consumers are overpaying vs what they really need.
Dropping from 95% to 57%, I don't think it's unreasonable that a time will come when Windows no longer has a majority.
I'm just back home from a long vacation. My computer while I was away was my $250 Pinebook Pro, it worked fine for everything that I wanted to do, that included a fair bit of development work.
Anyone who is a developer is likely comfortable with Linux. I feel most non-devs could probably make do with a basic distro, but there's plenty of small gotchas that we don't even give a second thought to.
Pine is clear who their audience is: "This is limited quantity production due to a severe global electronic components shortage. Please do not order the Pinebook Pro if you’re seeking a substitute for your X86 laptop, or are just curious and you’re ordering it with an intent to file a return/refund return request. These pre-orders are meant for enthusiasts familiar with the Arm architecture and interested in the PineBook Pro for this specific reason. Thank you."
However, this does underpin the point that ARM-based computers are extremely viable.
While I am also concerned about and need the extra performance, this point seems somewhat underappreciated in discussions about the ARM macbooks. Fan noise is so incredibly annoying and distracting and I think not having it improves productivity.
> Why is buying something 10x more powerful than you'd need at 2x the price "good value"?
So a MacBook Air is 2x the price of the very cheapest laptops. Trouble with the very cheapest laptops, though, is that they use the very cheapest chips; Intel is not making a lot of money on those (if they're Intel at all; these days a lot of that category is ARM Chromebooks). The MacBook Air's actual competition is more capable laptops in a broadly similar price range. These use more expensive, higher margin, usually Intel, chips, like the MBA did pre-M1. That has to be a concern for Intel.
I am a lifelong Mac user. My work laptop is Windows. Always has been, always will be. My employer probably buys more computers every week than I will buy in my entire life.
Nobody in the space gives a crap what laptops people buy for personal use once every 5-7 years.
Third world? I live in Western Europe and in my entire time before finishing college I saw a single Macbook in class. Among all people I personally know about 2-4 others use a Mac and we have a single Apple store in the entire country. Want a custom configuration from the officially licensed vendor? Have fun waiting 6 months.
When it came time almost two years ago to get a new laptop, I had a choice between a MacBook Pro for over $4000 or an Alienware that I could modify with 64gb memory and an extra SSD for half the price. Bare metal gaming with better performance and even though it doesn’t support Linux, it supports it better than Mac. Cost half as much. No futzing around with packages that don’t work on the M1, just good old AMD x86 goodness. The battery life sucks, but that rarely bothers me.
I’d love to buy an Apple, but it has never been priced competitively. My wife’s MacBook broke and it costs almost as much just to fix as to buy a new laptop.
I’m curious what those prices would look like today. In 2021 I bought a MBP M1 Max with 32 Gb RAM and a 500 Gb ssd for about 3k.
Looks like a max spec Alienware X14 R2 costs around 2,100 dollars. However the battery life seems to be the biggest drawback, most reviews say it lasts under 8 hours while my MBP lasts 12-16 depending on usage.
A similar priced mac > windows is ~2x the price, with no upgrade-ability options. I'm not so sure the price is right, it's expensive and they are making good money on it, but I agree does have some un-matchable features such as the battery life.
I work with both windows (with WSL) and mac OS, and there's nothing massively that sells me either way on the base OS, however I really wish tools like raycast existed in windows.
> the Apple M1 laptop is an incredible value, offers 10x the performance the average person will need, with good battery life, a better OS/ecosystem...and its priced right!
Ehh... Is it?
I use the M1 issued by my employer. It's an okay machine in terms of hardware, but I absolutely loathe MacOS. I certainly wouldn't pay for the M1 at the price they sell it, way over expensive toy.
I vastly prefer my personal machine, which is a 4 year old Dell with Linux Mint installed. A joy to use, powerful enough for everything I do (development, gaming), and has a GPU I can use to play with StableDiffusion.
The only thing where the M1 is the winner is in battery life... But 99% of the time I use my laptop on my desk anyway, power supply is no issue.
> and has a GPU I can use to play with StableDiffusion.
You can run stable diffusion faster on a M1 pro macbook pro than on a 1080ti 3900x desktop. DiffusionBee, I get .7s/iteration on a M1 using just 45w of power vs .98s/iteration on my 1080ti using over 5x that amount of power...
I could upgrade to a 4080 and I am sure it would destroy it, but I doubt your laptop has a full power 4080 anyway.
> But 99% of the time I use my laptop on my desk anyway,
As someone who works remote and has to travel for work, battery life is incredibly important to me, and the M1 is just leaps above its competition. I can write code and build all day without even plugging in on the M1 16"...
> You can run stable diffusion faster on a M1 pro macbook pro than on a 1080ti 3900x desktop. DiffusionBee, I get .7s/iteration on a M1 using just 45w of power vs .98s/iteration on my 1080ti using over 5x that amount of power...
At what cost?
> I could upgrade to a 4080 and I am sure it would destroy it, but I doubt your laptop has a full power 4080 anyway.
Mine has a GeForce 2060 RTX. Not in any way overpowered, but runs games and StableDiffusion fast enough.
That at a fraction of the cost of an M1 Pro too, and the added benefit of not having to use MacOS
The M1 MacBook Air is regularly on sale for $800 or less. I'm not sure how you can consider that an "over expensive toy". It's competitive and often beating Windows Laptops at that price. The M1 chip is 2 years old now, and still competitive.
I can understand hating the OS. But MacOS is my preferred OS, I really don't like Windows or Linux. MacOS just works. Linux and Windows always have some driver/kernel/etc. update that breaks something.
Like I said it is competitive with other computers in its class. The Surface Laptop 5 is $1000 and also comes with 8GB RAM and 256GB SSD. And MacOS Performas much better with 8GB than windows does.
The percentage of end-users who care about which language their OS is written in rounds to zero. Rust may add additional stability to the OS, but modern OS's feel pretty stable to the mom and pop users and "additional stability" is pretty far down the list of demands.
obviously, for Linux to ever see mass adoption it would need a drastic reformulation like Android
my point is, once Windows passes the event horizon, Intel needs SOMETHING or they are going to see 2/3 of their revenue stream go away...cloud vendors are busy pushing ARM in the datacenter now too...Intel is running out of markets
btw AMD is also screwed by all the same logic...x86 is on the endangered species list
Nonsense. Intel is one of the largest contributors to the Linux kernel, the biggest depending on what measure you count by[0]. They even have their own distro.
Moreover, much of their profit is from the server market, which tends to run Linux.
>While Intel's sales *beat its own expectations* in the first quarter, the company on Thursday posted the largest loss in its history as its margins plunged to a new low over the last several years.
Doesn’t the bathing come in the form of write offs and investments? Sales decline is never good. But it’s hard to say they have any good products Ive considered buying in the past two years or so. I actively right now wouldn’t consider a pc option if it came with intel.
It can make future results look better. If you don't mind reducing revenue now you may get in exchange more revenue in the future - and a more favourable baseline for growth calculations.
36% down YoY, but they beat investor Q1 consensus. Analysts predicted worse than this, so they've exceeded market expectations. Sapphire Rapids is only just starting to ramp sales, and the Emerald/Granite Rapids & Sierra Forest roadmaps are on track, which is the first time the Xeons have looked healthy in nearly 4 or 5 years. They trimmed the fat by selling off all businesses but their core ones, and are announcing IFS customers every other week. I'm not surprised investors are starting to move.
Can't edit my comment now, but I just looked at the earnings report and they're actually up $200m on what they forecasted, so that'll be why the stock is up.
“Business success contains the seeds of its own destruction. Success breeds complacency. Complacency breeds failure. Only the paranoid survive.”
Andy Grove
It seems they may be at an inflection point where they should move away from x86 and create the next big thing. Pretty much what Grove describes in his book.
I disagree. They need smaller, more efficient chips, they just fell behind.
Intel was the major driver of EUV in its inception. They funded ASML, only to sell their share later to keep the semis ecosystem open (oh the irony...).
They are making all the right moves now, but the semis cycle is ridiculously long. 5-6 years from planning to delivery.
x86 is what keeps them in the running for now, as there's so much software built on it, but if they don't stick to their roadmap and start leading with their products, that edge will also degrade.
I find it kind of amazing that Intel haven't found a way to profit during the global chip shortage. Look at food and energy over the past year or so: supermarket and oil & gas companies are posting record profits. They are riding the wave of price rises and supply constraints. What's different about CPUs?
They are very different types of goods... Food and energy are required by everyone, so when the supermarkets and energy companies decide to price gouge and profiteer (as they have been), people have few other options than just to pay the higher prices... People and businesses can mostly just delay buying new computers if they are too expensive.
There was huge purchasing of equipment at the start of the pandemic to enable work-from-home. Most of which is still current enough to not need replacing.
Meanwhile it's possible that the high food and energy prices are impacting the amount of money available to spend on tech goods ..
At the beginning of the chip shortage, Intel's CPU just weren't very good. AMD was beating them at performance and efficiency. Intel has caught up now, but their chips still use significantly more power than the current AMD chips.
Intel spent a lot of time and money developing their new Arc GPUs. They were delayed and they missed the crypto boom. Drivers and GPU performance were disappointing even after delaying the launch. This first generation of intel GPUs will be a huge loss for them.
Intel is less diversified today. They sold their LTE modem business to apple a few years ago. Intel exited the ARM chip business ~15 years ago, right before the smartphone took off.
Intel is now losing the server chip business. AMD is making competitive sever chips for a lower price. Larger companies are developing their own ARM-based server chips.
Intel is going to start using their foundries to manufacture chips for other customers. It remains to be seen if Intel can compete with TSMC.
To their credit, the gpu drivers have continued getting much much better, much faster.
Intel has long understood the value of drivers, has been one of the primary contributors & drivers of really good open source (albeit with some frustrating areas of sore neglect, like GMA500/Poulso) (But also rip the incredible http://01.org those folks were amazing at finding high value open source work to do!). Im glad to see making an already quite reasonable priced value GPU a lot better.
But yeah the timing is after the GPU craze. And Intel has been tweaking the same way-too-big big core for too too long & is uncompetitive. And they've been hurting to get a 7nm process (Intel 4) going. A lot of change is necessary.
One could conceivably assume that they're talking to Sony and Microsoft about putting their GPUs in their next consoles. I'm not sure Microsoft would bite for compatibility reasons but Sony seems ready to make another blunder.
It's possible, but AMD has supplied console APUs for two generations now because they are willing to sell them for cheap. Intel has never been interested in low-margin items, but they may need to be in the future. AMD has really targeted the gaming APU market. They make the chips for the Playstation, Xbox, and Steam deck. They also recently announced a new chip for gaming handhelds that Asus is using in their new ROG Ally device.
In 2020, Intel generated a record $35.4 billion cash from operations and $21.1 billion of free cash flow (FCF) and returned $19.8 billion to shareholders. They profited massively, and put their dividend up 5%.
They brought in Gelsinger and used that cash to build 2 new leading edge fans, and upgrade capacity on newer ones. Then the economic downturn happened, and sales dwindled. They're making less revenue but still funding the production of new fabs. That coincided with a weak roadmap. Sapphire Rapids is only just starting to ramp, after a nearly 3 year gap in their Xeon roadmap.
They missed the boat on 2 biggest growth areas: ARM (mobile) & GPU (gaming, crypto, AI), and screwed up x86 (AMD made some better chips / Apple walked away).
Screwing 1 of those up is a problem, screwing up all 3 simultaneously is near-bankrupting.
Their development model was 'tick / tock'[0] with the tick being smaller processes, increased frequency, lower-power. All of those are largely outsourced (or outsourceable) to a fab. The tock step was introduction of new microarchitectures. Intel has been historically hit or miss with this, the default is to take what they did and do more of it rather than come up with entirely new and sound designs. I would say for large established companies the greatest factor in these kinds of differences is how high the engineering culture goes up the reporting chain.
As a casual observer, Intel seems to have commitment issues. It's not like they haven't made or are not making attempts in some of these markets but they can seem half-hearted even when those ventures aren't mothballed or sold off outright.
I live in Columbus, and over the next few years I expect a lot of shakeup to occur as a result of the Intel fab(s) being built nearby. It's been interesting to see the turbulence around Intel of late, given what a huge investment we're expecting to see from them in Ohio (not to mention the huge investment the state will be making into Intel, via tax abatements and other incentives).
Let's hope it actually gets built and that the state isn't Foxconned like happened to Wisconsin. A company with rapidly declining profits does not give me a lot of confidence in starting or finishing big capitol projects even if they've committed to them. Nothing is set in stone until the last brick is laid and the factories are open.
Like I said, I'll believe it when I see it. At the end of the day capitalism rules supreme and trumps jingoistic nationalism. If it's cheaper for Apple and others to build chips overseas then all your devices will continue to have chips made there. I don't see any fundamental change to make labor in America competitive with overseas labor costs. Intel can build as many fancy fabs as they want in Ohio but they will be useless if the cost of living, cost of healthcare, median home price, rent, inflation, etc. continues to skyrocket. It's still too expensive for workers in America and building factories doesn't fix that problem.
> capitalism rules supreme and trumps jingoistic nationalism.
I'm not so sure about that. At an individual level sure but we're talking about the USG not a bunch of individuals at a national level. The government has made many decisions to subsidies/backstop local companies in the interest of nationalism.
There's no way it's cheapest for Boeing to make its airplanes in the US.
Janitors for the USG get security clearances.
Farmers (even by state governments) get subsidized to ensure local supply.
If something like the "Freedom Phone" actually used USA parts and was assembled in the USA I could see it becoming the governments defacto corporate phone. Sure, they might become a "Government Contractor" as opposed to an actually competitive company but they would still exist with real buildings and real employees doing real activies which is strictly what we're arguing about. The USG will backstop an inefficient business; Foxconn has no problem not building an inefficient factory.
Deciding to give Intel massive subsidies would require congress to agree and enact legislation, and that isn't happening with it being a politically fractured senate and house. It also isn't exactly sustainable to just keep funneling truckloads of money into Intel to be competitive and kind of pokes a giant hole in the idea that the free market is supreme at solving all problems. In the end subsidies just contribute to rising inflation (it's printing money). The fundamental problems have to be fixed.
> Deciding to give Intel massive subsidies would require congress to agree and enact legislation, and that isn't happening with it being a politically fractured senate and house.
Eh, sure congress has apparently only passed 1 bill so far this year [1] but you can go look up other split years (i.e. 2013 [2]) as a reference. They actually pass a lot of bills despite what people say and the first 4 months of the session are always slow.
Ohio is a R state so republicans won't bring up deficit issues (not that they actually care about them). Democrats (obviously) voted yes on the CHIPs act last year so its not like they're going to change their mind now. The remaining parties aren't a large enough coalition to matter.
> It also isn't exactly sustainable to just keep funneling truckloads of money into Intel to be competitive
Unless you're going to argue the USG is going to become insolvent prior to Intel finishing the fab in Ohio this is an irrelevant argument. The USG will be around longer than it takes to build a fab and as I brought up in my previous comment; plenty of companies are not competitive on a global scale but they still turn a healthy profit because the USG backstops them.
> In the end subsidies just contribute to rising inflation (it's printing money)
Well, you can subsidize something without printing money. I think printing money causing inflation has been trivial disproven though by the 2019 pandemic where governments printed money like mad and no inflation occurred until years after the fact. Unless of course you want to argue the free markets are so inefficient they took 2-3 years to correct for that money printing?
Yes, this is a major unsolved problem to my eye. The cost of living in Columbus - while nowhere near the craziness of coastal cities - is steadily rising. Real estate in surrounding cities is also going up, and it's hard for me to imagine how all those working the jobs that will supposedly be created by the fabs will afford to live near them.
One development in Ohio has been a proposal for a passenger rail line that would connect Cincinnati, Columbus, and Cleveland. I'm originally from Crestline - a small town which sits right on the rail line between Columbus and Cleveland, and which used to be a big railroad town. It's interesting to think of how my hometown might change if passenger rail were to actually happen (though I'm skeptical, as versions of this plan have been proposed for decades).
That said, there are some major problems to overcome if we wanted to start doing "state-scale" transportation like that - Columbus is a very car-centric city, and in the scenario where I would move home and ride the train to Columbus when I needed to go to the office, it's very unclear how I'd get from the station to wherever I'm going.
Furthermore, there was just a very public and catastrophic train derailment in Ohio, which can't do much good for public confidence in commuter trains.
Either way, the next decade or so in Ohio will certainly be interesting. Columbus has a burgeoning tech community with lots of VC investment, and while I'm not sure where I'd put the odds, it's not impossible that our little flyover state could become the proverbial "heartland tech hub".
Yeah I lived in Columbus and other parts of Ohio for 15 years or so and it strikes me as bizarre to try to bring industry back. Rubber, steel, automotive, etc. industries were absolutely massive in the 50's, 60's, 70's. But the corporate raiders of the 80s destroyed and sold off all the factories to replace them with cheaper offshore labor. The cost of living in Ohio only continued to go up and most cities in the state are filled with suburbs full of homes now approaching $500k or more. It's just not sustainable or sane to think some new class of factory workers can jump in and live there while being competitive on labor prices.
Yeah, upper middle class people are predictably out of touch with reality. The median household income in Franklin County is only $66k. Yes, stupid zoning made building non-McMansion houses difficult, but that doesn't mean the working class just magically disappeared.
> most cities in the state are filled with suburbs full of homes now approaching $500k or more.
I still get Zillow emails from my time living in the state. $500k is in the rearview mirror. A lot of new developments in the "nicer" suburbs are hitting the $1MM mark. $500k buys you a new home in the far, far suburbs. $250k gets you a 800 sqft 40s-50s ranch in a "crestfallen" neighborhood (and the price history on Zillow will show it as being sold for $60k 10 years ago).
Yeah I lived there when 60k could get you a home and 100k was, "woah there Mr. Mansion-haver" money. My mind boggles at spending 500k there now and barely getting a couple bedrooms. Nowhere in Ohio should it cost that much to have a home.
People in this thread are a bit clueless on the economics of it.
Many of Intel's existing fabs are in the US, and they've never really stopped building them here - particularly in Oregon and Arizona, and they continue to build there. The only difference is that now they're building on yet another site. The bulk of their fab process R&D is in the US, and they perfect the flow for a given process here and their international fabs then follow suit.
Why do they continue to build in the US? The CEO answered this over and over during his campaign for the CHIPS act: Running fabs is significantly more expensive than labor costs. They don't really save that much by building abroad.
When it comes to cheaper things like packaging, the labor costs do matter, and so they do that abroad.
So no, Ohio won't be foxconned. If Intel doesn't do well, Ohio will be dead, but they're not going to build elsewhere to save on costs. There isn't much to save.
Intel pulled the same trick in Washington State. After the plant was built and Intel got all the tax incentives and free money, they scaled the plant down to just package chips manufactured overseas.
Easy to say but what would that future be? The only thing that could replace Windows is Linux. Intel funds Linux but there are fundamental issues there that stop it taking over. They could try and develop their own OS and PCs, a sort of reverse Apple, but do they have the money, staff and ideas for it?
Intel and Google share surprisingly similar problems: an almost pathological inability to grow lines of business with new products because they kill them off too early. In Intel's case, it's even worse, they kill them off right before the market would have made them successful. They should own the "things with semiconductors" market, but instead financially engineered their way to focus only on higher profit core chip business which was good for the short term but has proved to be a festering cancer over the long.
Because of this approach they divested out of the memory business -- which has kept them out of the growing solid state business, screwed around with Optane, lost their modem business, have lost the GPU business at least once (who knows if they'll keep doing the ARC GPUs so maybe it'll increment up again), ARM chips, and numerous other bits and pieces over the years. Now they're losing their core business too and have nothing else to really grow and focus on.
Each of these is a multi-billion dollar market with core providers who dominate the market: Samsung, Broadcom, Nvidia, etc.
Now they're going to start selling off the bones of their business to compete with TSMC, a company with actual focus.
Once that goes away, what? Intel will contract TSMC to make their chips? I can see a financially focused MBA-type executive team seeing that as a way to eliminate expensive Fab R&D, construction and OEM liabilities -- put it on somebody else to deal with that!
Intel got out of memory when that became a complete commodity, similarly Intel sold their NAND business (now called Solidigm) a while ago to Hynix, because similarly NAND and SSDs are now commodities (in the sense of highly standardized, interchangeable product made by a large number of competing companies). It does make sense if you only wish to keep things where your company has unique advantages and doesn't need to compete solely on price in a commodity market.
You are absolutely correct in the history to my recollection. But I'd challenge that it was a good idea as having multiple product lines better creates opportunities for new and unique products in different spaces.
Optane is an example of where Intel tried to "fake it" but ended up producing a confused product that didn't leverage common industry knowledge in either the memory or solid state storage spaces. More towards a near future, Intel should be looking at compute-on-memory hardware, which promises to radically improve computation in a number of areas, and fundamentally would give them an "in" into neural network-like tech.
Similarly, the companies that do produce NAND and SSD technologies seem to have figured out how to make money in a commoditized and highly competitive market.
I agree! When memory became a commodity --- there are periods in markets when nobody confers distinction --- then maybe you keep your hand in to break even or small profit. But keep your eye on the next evolutionary step. Compute co-located with memory (non von Neumann) is that step. With the hw, there'd be an opening for a next gen OS to run it.
How did Optane represent faking it? It was and still is the fastest persistent storage ever made, no? I thought they killed it because of the difficulty of manufacturing it.
It died because the first generation of Optane devices failed to take the market by storm and instead of advancing the technology they simply exited the market, eating all of the prior investment.
Why they failed to take the market seems to be both a combination of lousy marketing (is it a drive, or are they RAM replacements?), selling only enterprise-class products, high prices, and competition technology that is both "good enough" and beats it in all of the other areas.
Spec-wise, a Samsung 2TB 990 Pro, is about as fast (sustained, IOPS, etc.), a 10th the price, and is coming from a manufacturer with a reputation of sustaining development and releasing newer, better products in that product-line. If the part fails, you're likely to not only replace it, but upgrade it for the same cost.
It's extra bitter to see Intel exit NAND because they largely invented & drove NVMe. The p3700 was the first mainstream NVMe drive, July 2014, & is still a beast, still pleasantly fast with stable & high iops, albeit lacking in raw bandwidth.
The before times were so awful. A sea of proprietary interfaces, dubious drivers. Intel identified a huge weak point in the ecosystem & created radical change where it was needed, taking something exotic & astronomically priced & building a stable better future, around what has always made PC so exceptional: standards. The flash industry was a mess, & Intel more than anyone made it real.
Intel is the poster child that if you want to run a giant business long term -- have independent, well-supported teams that deliver a profitable product on their own... and are insulated from being fucked with by anyone in the management chain.
Modern Intel would have shut down the Haifa team, which would have been an existential problem when Netburst reached the end of the line unexpectedly early.
At that scale, you need to keep teams/options around simply to retain expertise and hedge your tech bets.
Personally I think Intel lost the plot around the time they killed Atom. Atom launched as a garbage product, then they spent a decade refining it into an incredible combination of price, performance, and power consumption. Then they killed it and replaced it with processors that offered barely 10% better performance per watt at five times the MSRP and they wonder why the Intel tablet market is dead. It's dead because they strangled it in the crib by pretending Pentium Gold Y-series was an adequate successor to Cherry Trail.
Killing larrabee was the right move. That project was a joke architecturally - they got absolutely nothing right. Wrong problem, wrong solution, wrong time, poorly executed.
I think Intel's tablet ambitions and their larger Atom ambitions died because they tied them to Microsoft.
The Atom could have been bundled as an easy to integrate SoC for things like set top boxes, routers, and smart devices running Linux/Android The sort of uses that go for MediaTek or AllWinner ARM SoCs. Intel had all the peripheral IPs to make competitive offerings.
Instead they put the Windows millstone around the Atoms' neck and the chip sucked. Intel based tablets sucked because Windows was a dog shit OS for low end devices. The few Atom based Android tablets that got made were pretty decent devices.
I've got a bunch of Bay Trail devices lying around. The ones I've put Linux on end up being very useful little devices that run very well. My Windows Bay Trail tablets are just e-waste at this point. They're difficult to use at best and typically just slow pieces of crap.
The Atom lineage also lives also lives on as the E cores in current Intel desktop and laptop CPUs. Its also coming to the server market next year as the high density, all E core Sierra Forest.
Can you explain more? I also sense the Philips is a fading giant, but I struggle to a put a finger on it. Part of it might be that it is constantly spinning out divisions into separate companies.
They don’t just spin out divisions, they keep spinning out innovative and revenue generating divisions, leaving Philips with no ability to do anything, and with what they’re currently doing: selling the brand for Chinese white-labels to put on their product as they try to move up the chain.
Essentially Philips had spent 30 years selling the bones of the group, leaving just a skin-suit. Which they’re now selling as well.
They have been using TSMC for a few decades for many of their chips. The last ~5 years has been the cutting edge chips...but they still have a lot of old chips made by TSMC still (think ethernet controller chips, etc).
Is there really a “too late” here? GPU sales today are probably a small fraction of future AI (TPU?) chip sales. Seems like Intel should be racing in this direction no matter how long it takes them to catch up, no matter how unprofitable it is in the short term. But maybe there’s something fundamental that ensures they’ll never catch up?
> but instead financially engineered their way to focus only on higher profit
You explained the core of the problem that usually gets little blame. Successful companies get infiltrated by Wall Street MBA types who make decisions for the short term to manipulate the stock price because their own compensation structures help encourage that behavior. Companies, especially like Intel should be built for being resilient over decades instead of maximizing profits by the quarter.
>Successful companies get infiltrated by Wall Street MBA types who make decisions for the short term to manipulate the stock price
This really applies for Intel, but for Google, I think it's less so. I've been thinking you have to look at incentives and at a large corporation everyone is incentivized to make the most amount of money in the shortest amount of time. If you really want to spend years hammering away at something, it made more sense to just start your own company doing that where you would reap most of the benefits through equity. There is no reason for someone to spend years at Google on a project that is seemingly going nowhere; they would struggle to get promoted, and if it does work out, most of the value will go to existing shareholders.
Why not? Getting a project up is a struggle anyway. And it’s a trade off between the upside and downside risk. At Google there is little downside risk and upside is reasonable. One or two experienced and passionate engineers, hammering away, amplified by a big company can compete with a startup.
The downside risk is only marginally different. If you are "passionate" engineer at Google, going to startup means making 150k rather than 350k. Don't get me wrong, 200k is a lot, even in the bay area, but it's not like you are destitute. If it doesn't work out you can always return to Google. The upside is completely different.
Secondly, you are also at the mercy at your VPs, who got to that position by also maximizing short term results. A VC may not have the tools to force you to pivot if you are making ump-teen million dollars in year 1, but a VP may decide to kill your project to move more experienced engineers to another pet project. A passionate engineer at a pet project is an anomaly at a large corporation, who also has to fight the political winds of people trying to maximize their short term gains. And I don't think those people are wrong, it's the rational thing to do, especially when any equity upside will go largely to existing shareholders.
"Only the paranoid survive" applied fractally at all levels in Intel. This meant that there were very few "anointed successors" as people who had specialized knowledge regarded not sharing it as a survival technique.
This meant that as the technical people who built things got old and retired a lot of institutional knowledge simply walked out the door.
Technical excellence rests on training and sharing. Intel was the exact anathema of that.
I think Intel missed the boat when they refused to accept third party fabbing back when they were process node leaders. So instead that business went to TSMC and other competitors and made so much money that those companies were able to overtake Intel and are far more diversified so they don't live and die on the PC market.
At the time everybody was saying "of course Intel isn't going to fab AMD chips, that would be crazy", but in the end the crazy thing was giving up market dominance.
Seems like a theme with Intel. Back in the 80s Acorn was evaluating using Intel's 286, but Intel didn't want to work with them so they ended up making their own CPU, which directly led to ARM.
ARC GPUs are basically dead on launch. It's not competitive in any way and fundamentally broken upon release. Intel integrated graphics works because it's a value-add to an expensive component already necessary for many computing products.
ARC GPUs is basically Intel expecting the market to be the beta tester for a very expensive & inferior product.
have you seen GPU prices lately?
The intel a770 is quite a good deal, especially considering you get 16GB of vram when the competitors only offer this at a far higher price.
I kind of agree with you that the product has had issues at launch, a lot of them are resolved however.
Also, linux support is superb compared to nvidia.
In the next 1-5 years we are going to see a lot of big companies falling apart.
Small teams, with the help of tech/AI can now move so much faster than big teams. The communication and coordination overhead is starting to be too big of a burden, especially when you can have 1 person do the work of 30-100 (using AI).
I’m not sure that’s true given hardware as a limiting factor. Small teams have always been able to move faster in building things, they have fewer requirements and less to lose.
But small teams can’t buy expensive hardware up front, and modern AI is hardware constrained for many purposes. These companies will either end up renting hardware from cloud providers, or using AI APIs from cloud providers.
Hard to understand your comment in context of intel.
There are a handful fabs around the whole world who can even make those chips intel and tsmc can do.
No one will just replace any of them in any near capacity at all.
And the entry to this level of tech is not 'just a few millions' ist actually (just making a fab with people and processies of people who know how to do this) is 10-20 billions.
And if you would take any software company i also disagree: I was working in a small company, they didn't knew security really. I intriduced a ton of security. In comparision to the huge company i'm now: The developers still don't know much about security but the security teams of my company do.
In my opinion the gap will be even bigger because small companies (besides the cool and hip new elite startups) still have more of a traditional mindset while the big onces need to compete. They need to use AI and co and they have the financial background to create AI teams.
> when you can have 1 person do the work of 30-100 (using AI)
That’s quite the statement. I highly doubt that will be reality in the next 1-5yrs, especially for novel R&D work. Even more so for products in the physical world.
This is absolutely what happens when bean counters are promoted to CEO...they killed Intel a long time ago. Finally put an engineering guy in charge, but he is blowing smoke up investors asses...you can't dump billions in new fabs and technology while also losing billions per quarter.
Their foundry business is dead...they won't compete with Samsung and TSMC. Even GF found that was a losing proposition and they had a lead on Intel and endless pockets of cash.
It's interesting that every large American industrial corporation eventually succumbs to MBA fever.
In the 20th century, we saw the rise and fall of General Electric, Westinghouse, GM, RCA. Huge American corporations with products almost everyone interacted with almost every day. In the 1960's these were the corporations to beat. In the late 1960's, bean counter infiltration began. Of these companies, only GM isn't a hollow brand licensing shell, and it had to survive a government-maneuvered corporate bankruptcy to make it to 2009.
Something is going on with Xeon-W. Intel started "shipping" Ice Lake Xeon-W in June 2021 [0], yet today in April 2023 major PC OEM vendors such as Lenovo, [1] Dell, [2] HP [3] and Apple [4] are still shipping workstations with the older Skylake Xeons in them and none of them offer a workstation with Ice Lake Xeon-W. Either the chips are not available in volume or a large segment of the PC industry decided not to update their computers to use the new chips.
The case for Apple is sort of obvious, the Mac Pro has always been slow to rev and has always had slow sales. Apple has also been transitioning to their ARM CPUs so they're not going to double down on x86 on a single line of machines.
I'd imagine other manufacturers are in a similar boat. They put in orders for Skylake Xeons for machines that sell slowly. Pandemic sales went into laptops for home offices rather than $5k desktop workstations. Desktop workstations have never flown off the shelves but are flying off more slowly than normal.
Ryzen and i9 laptops and a beefy VPS have been good enough for many WFHers so no one is buying desktop workstations. Manufacturers aren't going to jump to new Xeon parts if they're sitting on old stock.
With the stunning 1-2% year on year gains in speed, when you're not losing 10-20% of your performance for mitigations, who sees a compelling reason to upgrade their computer? Laptop, workstation, or server?
Because the people moving big money on the stock market are not sitting around waiting until earning reports come out to make a move. They already knew intel was going to post a loss, this was not news to the market.
You're not wrong, but I think this misses the point.
Even retail knew they were estimated to post a loss. I'm guessing something in the earnings report was "less bad than expected" in order for the stock to be up 5% today?
Yes, the market will often move in the direction of the delta between expectations and actual results.
People often assume it "should" move in the direction of the overall sentiment (good earnings = up, bad earnings = down), but that is not at all true because earnings reports are a retroactive view of information people have already been trading on.
You don't want to go where the ball is right now. You want to go where you think it will be in the future.
I don't usually do sports analogies, but my kid recently started playing soccer and I've been trying to get her to intercept rather than chase people. Heh.
I think one of the biggest concerns right now (maybe always?) in the stock market is "things will stop being gangbusters, the economy will slow, growth will slow, and the ball with stop being projected to be in front of you/forward at any scale/pace"?
It's not actually vital for Intel as a company to survive? As long as the foundaries and designs survive. Somebody else will take the business away from them.
Intel is well underway on the same journey that IBM took decades ago. That is the transition from a Computer Engineering company to a Financial Engineering company. This never works well. Never, ever, let an finance or sales person near the top spot at a technology company.
291 comments
[ 3.0 ms ] story [ 276 ms ] threadthey're still tied to Windows which is increasingly weaker and more rotted over time, losing market share every year for a decade (Microsoft doesn't even seem to care afaict)
the Apple M1 laptop is an incredible value, offers 10x the performance the average person will need, with good battery life, a better OS/ecosystem...and its priced right!
Intel needs to seriously look at Linux...once Windows enters into a death spiral in a few years (seems inevitable now and Microsoft seems to not care as long as Azure keeps winning), what else will they have?
start now with serious price cuts....I don't need to replce my Linux laptops, but would for the right price
Cheapest one you can currently buy is the Air. That starts at 1k USD pretax. That's not within reach for a large majority of the worlds population. Not even within reach for a lot people in richer western countries.
Most people need a Webbrowser that can do Email, Streaming, and some social media.
What makes you think windows is in a death spiral? They still have a good 75% market share and while that fluctuates, it doesn't seem to really drastically decrease.
Of what? Most of what I see normal consumers doing is in Android, typically on ARM. I think my observations are a little biased, but not a lot. According to statcounter:
Android 41.56%
Windows 28.29%
iOS 17.29%
OS X 7%
Unknown 3.05%
Chrome OS 1.32%
Critically, it's no longer a monopoly; there's a lot less lock-in than there once was. Even Microsoft Office is gradually moving to the PWA model. Most servers and embedded systems I see run some variant of Linux as well. The market Microsoft has 75% share in seems to increasingly by a niche market within a much larger ecosystem. Microsoft seems to have pivoted, and as someone pointed out, no longer cares too much.
However, the Wintel monopoly is dead. If Intel is tied to Windows, they'll increasingly be swallowed alive. TSMC manufactures most of those ARM chips. There's no way a smaller player can keep up in fabs and manufacturing technology.
Comparing cell phone OS to a desktop OS is a bit silly. Nobody in the corporate world is doing work on Android.
If you'd like to see how well trying to focus on that niche works, have a look at DEC Alpha, MIPS, PA-RISC, Itanium, Sparc, PowerPC, etc.
At the end of the day, it comes down to manufacturing volume. Volume leads to investment in next-generation technologies. More investment means, at some point, competitors can't keep up. Manufacturing doesn't care too much about who the ultimate end user is.
Even big players like Microsoft are forced to offer their software and services on non-Windows platforms (like mobile and tablets!) if they want to remain relevant.
So counting desktop only paints a false picture, wrt where users are and where computing is happening.
And that last bit is what is relevant from a Intel POV. An increasing amount of compute is now done outside the Intel-chip arena of desktops, like it used to be before.
These industries won't buy macs as they are too expensive to maintain and to integrate into their existing active directory domains. (With a few exceptions for jobs where the Mac is the optimal device like graphic design)
No idea about percentage but windows won't disappear soon from those.
The android market share is mainly mobile devices and Microsoft gave that market up a few years ago, focusing on azure and office. Here again, focusing the corporate/gouv world.
The corporate/gouv world has many windows only app and processes that would cost a fortune to replace.
Windows is here to stay, just like mainframes.
I generally agree with your point though but I am seeing that as more and more of office work moves to living within a web browser, it is becoming more hardware agnostic.
No. It's reality. Mainframes are here to stay. IBM continues to develop and sell plenty of mainframes. Several other manufacturers do as well.
https://en.wikipedia.org/wiki/IBM_Z#IBM_z16
Where I work has applications dating back decades with complex data analysis algorithms developed by very smart people. Strict backwards-compatibility is critical. We have mainframes.
There are many other reasons to run mainframes. For example, mainframes offer pretty incredible features with regards to redundancy, and making everything hot-swappable in case of failure without interrupting workloads. That's especially useful in places like the financial industry. There, a single failure or error can cost much more than a mainframe.
They're a niche, but a big enough one to support a multi-billion dollar business. It shrunk for many years, but it's currently even growing.
Don't get me wrong, I'm glad I don't have to deal with it anymore, but mainframes are definitely still around.
Back in the day, they cost a fortune to set up and replacing them is a high risk - low value for many CEO's.
In Québec, the SAAQ (provincial DIV) decided to get rid of the mainframe by replacing it with SAP and with the amount of data they had, the migration took them a month of downtime.
Then they completely mismanaged the startup of the new system and a lot of people complained. The CEO was fired. https://montreal.ctvnews.ca/head-of-saaq-loses-his-job-eric-...
I'm sure that Windows Server will gradually be replaced by containerized app running in some cloud but windows desktop will take a long time to be replaced, especially because of office.
I'm not convinced. About ten years ago one could say exactly the same thing about Macs companies located outside of California. I only had my first work Mac in 2012, when I convinced my boss to let inherit one from a graphic designer. Since then, I've worked in two extremely traditional corporates that gave me the choice to use macOS. And plenty of executives were/are using macOS too.
In the meantime things got actually easier than 10 years ago: web apps became prevalent and even Microsoft joined the party.
but if it continues to shrink (and it almost certainly will), then Intel (and AMD) are in big trouble
Intel can't exist servicing a niche
IBM isn't trying to float a multibillion market cap solely on the back of their mainframe CPUs...
and of course let's not forget that Microsoft is also continuing to hint at future Windows-on-ARM...the first try failed but I have heard more talk...
https://www.bestbuy.com/site/macbook-air-13-3-laptop-apple-m...
I feel like the Chromebook on the lower end is great for a lot of uses/people and the Air is an easily achievable next entry-level step up.
Looking at the best selling laptops at Microcenter, they're $600, $700, $700, $1100, $800, $800, $750, $900... Having an M1 Air slotted in at $800 seems right in-line with the market.
BTW, the cheapest M1 Air (8 GB RAM) in Austria is 1200 EUR.
If you can stick with a terminal or some lightweight editors, 8GB might be okay for you, but I feel like 8GB has been pushing it for at least 5 or more years now.
There was a lot of discussion of this at the time the M1s first came out, with loads of people (who had never touched one) confidently declaring that 8GB of RAM was far too little...and then nearly everyone who actually bought an 8GB M1 Mac saying "yeah, it might not handle heavy loads well, but that's what the Pro machines are for—in normal use, it has no problem at all with 8GB."
And then a bunch more people talking in a fair amount of detail (that kind of went over my head) about what Apple did to improve memory management with the M1 so that this would be possible.
So don't assume that 8GB on an Intel Mac is the same as 8GB on an Apple Silicon Mac.
Now, it does sound like you were trying to run a whole bunch of rather heavy stuff on it—y'know, the stuff that one might call "pro", and possibly even buy a "pro" machine for? The kind of thing where if you're buying the absolute bottom of the line, you might be able to expect that it wouldn't work well.
I own a base Macbook Pro 14", 16GB RAM. If anything, it feels like less than 16GB (maybe because the shared video memory or macos bloat) saved by the fast SSD that's inevitably going to fail not too far and turn the laptop into e-waste.
I feel like every single talk point of the early review cycle was a huge exaggeration:
1. battery life just browsing the web is no more than 4/5 hours. Like any other windows laptop.
2. Performance is really nothing special really. Things like maximizing/minimizing a youtube video are choppy and lag for 2/3 seconds.
I mean, it's still the best value in the market for the screen, keyboard, chassis, speakers combo but Apple Silicon is extremely overrated.
That's nice, but I can buy a much cheaper non-"pro" computer with better specs that can handle a basic workload that involves a browser, IDE and maybe Docker. Like I've run all of that on entry level machines without a problem.
I bought my laptop in 2015. This underscores your point: Buy something upgradable with decent specs and you'll get a lot more bang for your buck.
Because Apple will forcibly kill it after that with incompatible software updates. Windows machines keep going until they physically break.
But what everyone here is missing is that the consumer desktop/laptop market is dying itself. Soon it will only be for enterprise uses and gaming. Most people do all their computing through a phone or tablet. I know people who don't own a laptop.
That's interesting, as I just this month replaced the two kids' Dell 9020s with new PCs because the 9020s didn't support TPM 2.0, required by Windows 11.
https://support.microsoft.com/en-us/windows/enable-tpm-2-0-o...
"Most PCs that have shipped in the last 5 years are capable of running Trusted Platform Module version 2.0 (TPM 2.0). TPM 2.0 is required to run Windows 11, as an important building block for security-related features. TPM 2.0 is used in Windows 11 for a number of features, including Windows Hello for identity protection and BitLocker for data protection."
The 9 year old Dells were running just fine on Windows 10 (same as the 9 year old Air was on Catalina), but upgrading to the latest OS in both cases required a purchase.
That was the actual driver, though this thread is about OS vendors forcing changes in order to run the latest OS, which is why I focused on the Win11 TPM2 requirements and mentioned the 9 year old Air running a still-patched Catalina.
To the point of this sub-thread, it's still an OS vendor forcing a purchase to run the latest OS. 9 year old Macbook running Catalina and 9 year old Dell running Win10, both unable to update to the latest OS seems very close the same to me.
Most windows machines I had worked for around 3 years. 5 max.
This is demonstrably false. My wife continues to use my hand-me-down MacBook Air I bought in 2011. It was supported officially by the macOS until the most recent release, so she’s only one OS revision behind. I also recently retired a 2009 Mac Mini that I used 24/7 as a home server, only because the power supply failed. I used the G4 Mac I bought in 1999 as my daily driver (with some upgrades) until 2009.
A decade or more of use is an incredible value, and I’ve done that multiple times with my Macs. Few of my PCs have had a useful life approaching that length of time (the closest was an Atom machine I ran for 7 years).
Um, I have a 2012 MacBook Air (11"), that I use for my low-end testing, and as a Zoom host. Works fine. Runs lots of software.
It has to run Catalina (Mac OS 10.15), but Apple still pushes out security patches to it (I think we're up to 10.15.7).
I wonder if it was a recall. I didn't speak German that well back then so it might have gotten lost in translation. I recently also had issues with my Airpods Pro and they replaced for free, but they told me it was a recall.
It was covered outside the normal warranty because of the swelling issue. Mine swelled to the point where the trackpad was slightly unreliable.
Maybe this is more of an indictment rather than a defense of Apple, but one machine I keep on an OS that's over 5 years old because it has the one version of Apple's Photos that didn't yet screw up the retouch tool. (And I use that tool exhaustively in retouching old family photos I scan in.)
which isn't good for Intel!
cpus are expensive to develop...
I bought a low end Mac in 2015 with the same 8GB/256GB specs, 8GB wasn't enough even for browsing/chat/media/etc at the same time, and the storage space gets eaten up by the macOS base system and apps.
I doubt 80%, maybe even 90%, of the population needs anything more than that.
My mom uses a cheap Dell laptop that is 10 years old. I bought at the time for around 350 dollars is memory serves (it was in a different currency, so it is hard to do the conversion).
She is non-technical, and it absolutely suits her needs to this day. Only thing I did in terms of maintenance was buying a new battery.
The cheapest Mac just won't get the same cost/lifespan ratio.
Keep in mind that the vast majority of users are non-technical.
> the Air is an easily achievable next entry-level step up.
Costs $1575 in my country, and significantly worse spec than laptops in the same price range
Some people simply do not care about tech to a point where almmost any price is too high. I just had to persuade a relative to spend $250 to replace a 16+ years old Intel Core Duo with an HDD. It takes 2 minutes to open a browser window on that thing and she didn't even consider that a problem. The replacement will probably last until it refuses to boot.
Nothing ever comes to the quality at that price, so even if you can buy much cheaper laptops those suffer in quality and performance. This means the value per dollar is not any better than MacBook Air, unless you are extremely short on dollars.
Also the re-sell value of Apple products is good and combined with the durability of Apple laptops(this is where the repairability folks fail to convince most people because for most people Apple devices don't break down, they become obsolete after extensive use) the money spent on Macs turns to be an investment. The "Apple premium" you pay when you buy an iDevice is actually a recoverable when you sell it.
There's also plenty of people who use their laptop until it's dead. Infact, I'd argue it's probably more economical to purchase a repairable laptop and replace parts as needed.
Your advice boils down to just get richer and don't be poor.
but that hurts Intel too...their profit margins are in their high-end CPUs
people keep talking about the cost of Macs but they're cheaper than most Windows laptops...
go look at the HP, Lenovo sites...they're not steering people toward $400 laptops
Of course companies push the items with the highest margins. They want to make money after all, but they still have plenty of options in the sub 600 category that will last 5+ years under normal use without issue. There's obviously a market for low priced laptops, otherwise so many companies wouldn't build them. And it doesn't hurt Intel. If the choice for them is to sell a low margin CPU or none at all, they're gonna sell the low margin CPU.
You can get a used thinkpad with an 8th gen intel and 16gb ram for $200 maybe.
Sure the cpu will be slower but with 8gb ram on a mac you're not going to be doing much anyway.
Bad deal overall. At that price I would prefer a Chromebook.
I wonder why people buy those about Air heavy, about Air performing premium and expensive laptops then.
I'm glad you are happy with it, for 500 bucks seems like a fair deal, but this is nothing like Macbook Air.
I run an old (I think maybe 4 or 5 years) thinkpad with 64GB RAM. I see often these claims of superior performance of new macs, and it leaves me scratching my head. I do not have (almost) any workflows where the performance is not sufficient. For the ones where performance is not sufficient, I am happy to turn on my desktop. So my question is, what are the workflows people have benefit from the superior performance of new macs and at the same time do not justify working on a beefy desktop?
Beyond that, it depends on what you do.
The biggest threat to Windows isn’t Apple and Linux, it’s mobile phones. People aren’t buying computers like they used to. Most day to day things can be done on a smartphone now.
I see this first hand in my own family. My wife and mom both use their phones for pretty much everything. Email, web browsing, etc. If my wife needs a bigger screen for something like a yoga video, she usually uses a TV in the house which has a Roku, chrome cast, or similar attached.
Get the MacMini. The low end M2 Mac mini is $599.00 USD right now, although the one I run for my desktop is the more expensive one at $2,199 with 32GB, more cores, more gpu and more drive space.
Their gross margin is between 40 and 50%. Call it 50% to be generous and make the math easy. If their selling price today is 6X and they sell U units, their gross profit is 3 * X * U. If they cut the selling price to 4X and sell 2U units, the resulting profit is around 2 * X * U. If they were somehow able to move 3 times as many units with a 33% price reduction, their gross profits would stay about the same as today.
What would a 33% reduction in the price of the Intel components in your laptop mean at retail? A 10% reduction in the shelf price? 15% maximum?
Macs have one big advantage - efficiency, which translates to long battery life and more importantly for me personally - NOT OVERHEATING.
However, porting software to apple silicon results in inefficiencies, and for developers it gets a bit trickier.
I have a M1 Max with 65GB ram and I love it, but the moment Intel comes out with chip that doesn't overheat and is as performant, I'll be buying an Intel based laptop (granted, I will have to wait for longer).
In a perfect world, Apple would get back to offering Intel as a CPU option when Intel comes back to the leading edge. You'd get the Mac OS, x86 and performance in a well built laptop - a dream!
Does it have to be Intel, or would you consider AMD?
Yes, if you look back even just to Zen / Zen+ laptop chips, they were not impressive, but Zen 2 and newer based chips (Ryzen 4000 or higher) are performant, efficient, and do not overheat. There are some crazy 7045HX chips that are basically desktop chips that you wouldn't want in your laptop, but the 7040HS line is exactly what you were waiting for (as well as the 6000U and 6000HS line). Unless you need Intel over AMD for some reason?
Given them another node shrink and they'll be even better.
Of course they care.
Windows, at one point, had 95% market share. Do you believe that to be sustainable? Do we want that? No, in both cases. They are still dominant in desktop computing.
>offers 10x the performance the average person will need
Why is buying something 10x more powerful than you'd need at 2x the price "good value"?
>a better OS/ecosystem
This is debatable. I don't find MacOS to be all that great, personally.
>once Windows enters into a death spiral in a few years
You should have a look at Windows' desktop market share and get back to us. They still own the space.
I suspect most users are not the same as the HM crowd.
I would argue in the same way that buying an expensive pair of boots that last longer than a cheap pair to be good value.
> You should have a look at Windows' desktop market share and get back to us. They still own the space.
What is the trend line and how much is that space growing by relative to other devices like mobile phones and tablets?
While a fair argument, I think its also fair to counter that the conditions in which you may use your boots don't arbitrarily change and get less efficient and more difficult over time. Sure maybe sometimes you'll be hiking through mud but that will dry out and get back to normal, whereas software might turn into quicksand as hardware gets older. Future proofing your hardware is effective, but only to a point.
Ah yes the dwindling, profit starved desktop market
Yes they're still at a high percentage, of an ever smaller market
Not going the way of the vinyl record yet, but who knows?
If they care they sure don't act like it. Microsoft is clearly squeezing Windows users for whatever revenue is left. Nothing else explains the user hostility, backwards compatibility breaks and unending advertising and data collection.
They like that candy crush is pre-installed and that Windows shows them random spam news articles that they find interesting.
I hate it and use a Mac, but I suspect most users aren't like you and me.
They don't know the difference.
Take a look around and most people under 30 seem to have a MacBook. At least here in my area of the United States.
I have 9 nieces and nephews, plus some neighbor friends that have kids. Of them, maybe 2 or 3 use iPads. The ones that have laptops (other than school provided) have them for PC gaming, and they run Windows. I guess I'm in a different area of the U.S. from you, based on my anecdotal data.
nothing to "administer"
the platform is curated
it does what they need
it runs on battery for a long time
Intel (and AMD) needs a good consumer OS to keep it relevant in the next decades. It's what keeps the DEVELOPERS DEVELOPERS DEVELOPERS on x86, and therefore use x86 server architectures. Either that, or it is a commodity ARM business.
Microsoft has never been up to the task of a good consumer OS. Intel always should have had multiple irons in the fire, and with Linux they had a built-in weaponized competing consumer OS. They delegated all leadership to Microsoft rather than keep irons in the fire. WHich is ... ok ... when MS is the big monopoly on the block. That doesn't exist anymore.
If Apple pre-iPhone produced OSX out of probably less than 100 or so engineers... Intel could have done that with 1/1,000th of their quarterly revenue. Or less! It's just spitshining an existing, working kernel, and adding a bunch of testing for the cacaphonous PC hardware/peripheral market that they were already at the nexus of.
Apple had revenues of 6 billion in 2001, the first year of OSX. Intel make 26.5 billion in "a bleak financial year".
It's not too late! It's still peanuts to do a spitpolish of Linux and do hardware testing/driver development for desktop linux! AMD or Intel (or a shared foundation) could do it! It could probably happen within a month. Partner or acquire ubuntu, or Suse, or ... anyone.
This isn't just insurance. Intel has had DECADES of frustration getting Windows to fully use/support its hardware and CPU features. A lot of the time Windows would hold people hostage by waiting a major OS release (like for USB) before supporting something.
Not with Linux. Boom, your stuff is supported as soon as you integrate it into your distro.
This is the path to a key thing that Intel doesn't realize it lost: relevance. Microsoft has decided Windows isn't really relevant. That was Intel's only path to relevance.
All the things they wanted to stay relevant in: x86 on phone, SSDs in PCs, Optane, etc. The path to that relevance was having a usable consumer OS. That you had good control over. You get your hardware demo'd and supported at maximum speed.
Don't like Linux? Well fine, pick up OpenBSD, or whatever remnant of BeOS is out there. RedoxOS? whatever.
Gelsinger probably is following a path of "get engineering" fixed. Well, ok, good strategy given how bad the MBAs fucked up everything. But Intel isn't the big daddy and that doesn't really solve the problems, it just mitigates the biggest bleeding.
Intel needs SOFTWARE RELEVANCE. Intel does not have that anymore. Intel still has the resources and marketing power to achieve that. THE WINDOW IS CLOSING FAST. Regain software relevance.
Just do SOMETHING.
Intel is in real serious trouble, and so is AMD
Intel has a lot more denialism and inertia in their management to fully embrace / not undermine a legit ARM product.
[1] https://www.intel.com/content/www/us/en/events/accelerated.h...
M1 performance won't be more than you need for long. When the dev community finishes transitioning to M1/2, that will be the performance baseline for lazy websites and electron apps. Intel laptops will have their CPUs pegged to 100% constantly when running more than one tab in Chrome. We're already seeing this at my company. Intel performance regressions are discovered the last minute before release, because everyone is using M1. And the solution is usually just to turn off the offending feature for Intel Macs. It's legacy.
Who buys desktops anymore? Sure, it’s a market, but not the one it used to be.
https://gs.statcounter.com/os-market-share/desktop/worldwide
You can commonly find Macbook Airs at $800, and I doubt many are buying sub-$400 laptops, unless they are going with low-end Chromebooks, in which the Chromebook vs Windows is a reverse spin on whether consumers are overpaying vs what they really need.
Dropping from 95% to 57%, I don't think it's unreasonable that a time will come when Windows no longer has a majority.
Pine is clear who their audience is: "This is limited quantity production due to a severe global electronic components shortage. Please do not order the Pinebook Pro if you’re seeking a substitute for your X86 laptop, or are just curious and you’re ordering it with an intent to file a return/refund return request. These pre-orders are meant for enthusiasts familiar with the Arm architecture and interested in the PineBook Pro for this specific reason. Thank you."
However, this does underpin the point that ARM-based computers are extremely viable.
While I am also concerned about and need the extra performance, this point seems somewhat underappreciated in discussions about the ARM macbooks. Fan noise is so incredibly annoying and distracting and I think not having it improves productivity.
So a MacBook Air is 2x the price of the very cheapest laptops. Trouble with the very cheapest laptops, though, is that they use the very cheapest chips; Intel is not making a lot of money on those (if they're Intel at all; these days a lot of that category is ARM Chromebooks). The MacBook Air's actual competition is more capable laptops in a broadly similar price range. These use more expensive, higher margin, usually Intel, chips, like the MBA did pre-M1. That has to be a concern for Intel.
E.g. m2 air with bare minimum upgrades is 2000€. That is much more than double. So sure it is good, but at that price point that is the bare minimum.
Sticker shock looking at Apple product prices in Japan.
for the last ten years Microsoft has pushed all of its effort into Azure, which is coincidentally based on Linux (!)
accept it, they gave up on Windows. Windows 11 is my proof...a lame re-skinning which can't even displace Windows 10
> You should have a look at Windows' desktop market share and get back to us. They still own the space.
no one under 25 uses Windows! Go look at any college classroom and get back to us! Windows has been in decline for ten years, the stats aren't a fluke
seriously...just walk into any college classroom in America...
go see what ecosystem anyone under 25 is invested in...now tell me what the future looks like
Nobody in the space gives a crap what laptops people buy for personal use once every 5-7 years.
maybe they don't, but Intel should
btw my employer (Fortune 100 co) has switched to deploying Macs and will probably deploy close to 100k, all the Windows laptops are being timed out
America isn't the world. Walk into any classroom outside the US and you'll see something very different.
go look at high-end Lenovo laptops with Intels' best CPUs...$3k+++
Intel and AMD won't last long selling cheap CPUs to a shrinking Windows market overseas
I can't see any shrinking Windows market here.
I’d love to buy an Apple, but it has never been priced competitively. My wife’s MacBook broke and it costs almost as much just to fix as to buy a new laptop.
Looks like a max spec Alienware X14 R2 costs around 2,100 dollars. However the battery life seems to be the biggest drawback, most reviews say it lasts under 8 hours while my MBP lasts 12-16 depending on usage.
I work with both windows (with WSL) and mac OS, and there's nothing massively that sells me either way on the base OS, however I really wish tools like raycast existed in windows.
Does Linux need Intel though? ARM Linux systems are surprisingly nice. Asahi Linux is probably going to be the best Linux laptop experience.
What should Intel be doing with Linux that they haven't already been doing for like 20 years?
https://clearlinux.org/about
Ehh... Is it?
I use the M1 issued by my employer. It's an okay machine in terms of hardware, but I absolutely loathe MacOS. I certainly wouldn't pay for the M1 at the price they sell it, way over expensive toy.
I vastly prefer my personal machine, which is a 4 year old Dell with Linux Mint installed. A joy to use, powerful enough for everything I do (development, gaming), and has a GPU I can use to play with StableDiffusion.
The only thing where the M1 is the winner is in battery life... But 99% of the time I use my laptop on my desk anyway, power supply is no issue.
You can run stable diffusion faster on a M1 pro macbook pro than on a 1080ti 3900x desktop. DiffusionBee, I get .7s/iteration on a M1 using just 45w of power vs .98s/iteration on my 1080ti using over 5x that amount of power...
I could upgrade to a 4080 and I am sure it would destroy it, but I doubt your laptop has a full power 4080 anyway.
> But 99% of the time I use my laptop on my desk anyway,
As someone who works remote and has to travel for work, battery life is incredibly important to me, and the M1 is just leaps above its competition. I can write code and build all day without even plugging in on the M1 16"...
At what cost?
> I could upgrade to a 4080 and I am sure it would destroy it, but I doubt your laptop has a full power 4080 anyway.
Mine has a GeForce 2060 RTX. Not in any way overpowered, but runs games and StableDiffusion fast enough.
That at a fraction of the cost of an M1 Pro too, and the added benefit of not having to use MacOS
I can understand hating the OS. But MacOS is my preferred OS, I really don't like Windows or Linux. MacOS just works. Linux and Windows always have some driver/kernel/etc. update that breaks something.
It comes with 8GB of RAM and 256GB of disk.
Do you really need to ask why I call it an over expensive toy?
I wouldn't buy either of those.
I'm not sure, on the contrary I think in a few years Windows will be rewritten in Rust and entering a virtuous spiral. We'll see!
The biggest problem with linux is that it's made by people who like linux.
my point is, once Windows passes the event horizon, Intel needs SOMETHING or they are going to see 2/3 of their revenue stream go away...cloud vendors are busy pushing ARM in the datacenter now too...Intel is running out of markets
btw AMD is also screwed by all the same logic...x86 is on the endangered species list
Nonsense. Intel is one of the largest contributors to the Linux kernel, the biggest depending on what measure you count by[0]. They even have their own distro.
Moreover, much of their profit is from the server market, which tends to run Linux.
>the Apple M1 laptop
Is very expensive and is running MacOS.
[0] https://truelist.co/blog/linux-statistics/
https://news.itsfoss.com/huawei-kernel-contribution/
https://thenewstack.io/contributes-linux-kernel/
> Intel says margins will recover in second half of 2023, shares rise
stock market movements are not to be trusted a period less than a quarter at the very least, better yet an year, imho.
>While Intel's sales *beat its own expectations* in the first quarter, the company on Thursday posted the largest loss in its history as its margins plunged to a new low over the last several years.
The next couple of years will be make or break for them, so Im expecting to see some good things coming out of their pipeline.
It can make future results look better. If you don't mind reducing revenue now you may get in exchange more revenue in the future - and a more favourable baseline for growth calculations.
[1] https://en.wikipedia.org/wiki/Andrew_Grove
Intel was the major driver of EUV in its inception. They funded ASML, only to sell their share later to keep the semis ecosystem open (oh the irony...).
They are making all the right moves now, but the semis cycle is ridiculously long. 5-6 years from planning to delivery.
x86 is what keeps them in the running for now, as there's so much software built on it, but if they don't stick to their roadmap and start leading with their products, that edge will also degrade.
Meanwhile it's possible that the high food and energy prices are impacting the amount of money available to spend on tech goods ..
Intel spent a lot of time and money developing their new Arc GPUs. They were delayed and they missed the crypto boom. Drivers and GPU performance were disappointing even after delaying the launch. This first generation of intel GPUs will be a huge loss for them.
Intel is less diversified today. They sold their LTE modem business to apple a few years ago. Intel exited the ARM chip business ~15 years ago, right before the smartphone took off.
Intel is now losing the server chip business. AMD is making competitive sever chips for a lower price. Larger companies are developing their own ARM-based server chips.
Intel is going to start using their foundries to manufacture chips for other customers. It remains to be seen if Intel can compete with TSMC.
Intel has long understood the value of drivers, has been one of the primary contributors & drivers of really good open source (albeit with some frustrating areas of sore neglect, like GMA500/Poulso) (But also rip the incredible http://01.org those folks were amazing at finding high value open source work to do!). Im glad to see making an already quite reasonable priced value GPU a lot better.
But yeah the timing is after the GPU craze. And Intel has been tweaking the same way-too-big big core for too too long & is uncompetitive. And they've been hurting to get a 7nm process (Intel 4) going. A lot of change is necessary.
To be fair to Intel, some of this was legitimately overwhelming bad luck - their GPU driver team for Arc was based out of Russia.
They brought in Gelsinger and used that cash to build 2 new leading edge fans, and upgrade capacity on newer ones. Then the economic downturn happened, and sales dwindled. They're making less revenue but still funding the production of new fabs. That coincided with a weak roadmap. Sapphire Rapids is only just starting to ramp, after a nearly 3 year gap in their Xeon roadmap.
Screwing 1 of those up is a problem, screwing up all 3 simultaneously is near-bankrupting.
[0] https://en.wikipedia.org/wiki/Tick%E2%80%93tock_model
However they seem to have missed a number of tick cycles as of late, so their efficiency gap to ARM got worse and worse.
This means Intel didn't lose Apple just because of execution because Apple could have trivially gone to AMD.
I'm not so sure about that. At an individual level sure but we're talking about the USG not a bunch of individuals at a national level. The government has made many decisions to subsidies/backstop local companies in the interest of nationalism.
There's no way it's cheapest for Boeing to make its airplanes in the US.
Janitors for the USG get security clearances.
Farmers (even by state governments) get subsidized to ensure local supply.
If something like the "Freedom Phone" actually used USA parts and was assembled in the USA I could see it becoming the governments defacto corporate phone. Sure, they might become a "Government Contractor" as opposed to an actually competitive company but they would still exist with real buildings and real employees doing real activies which is strictly what we're arguing about. The USG will backstop an inefficient business; Foxconn has no problem not building an inefficient factory.
Eh, sure congress has apparently only passed 1 bill so far this year [1] but you can go look up other split years (i.e. 2013 [2]) as a reference. They actually pass a lot of bills despite what people say and the first 4 months of the session are always slow.
Ohio is a R state so republicans won't bring up deficit issues (not that they actually care about them). Democrats (obviously) voted yes on the CHIPs act last year so its not like they're going to change their mind now. The remaining parties aren't a large enough coalition to matter.
> It also isn't exactly sustainable to just keep funneling truckloads of money into Intel to be competitive
Unless you're going to argue the USG is going to become insolvent prior to Intel finishing the fab in Ohio this is an irrelevant argument. The USG will be around longer than it takes to build a fab and as I brought up in my previous comment; plenty of companies are not competitive on a global scale but they still turn a healthy profit because the USG backstops them.
> In the end subsidies just contribute to rising inflation (it's printing money)
Well, you can subsidize something without printing money. I think printing money causing inflation has been trivial disproven though by the 2019 pandemic where governments printed money like mad and no inflation occurred until years after the fact. Unless of course you want to argue the free markets are so inefficient they took 2-3 years to correct for that money printing?
[1]: https://legiscan.com/US/legislation/2023?type=bill&status=pa... [2]: https://legiscan.com/US/legislation/2013?sort=asc&order=Last...
One development in Ohio has been a proposal for a passenger rail line that would connect Cincinnati, Columbus, and Cleveland. I'm originally from Crestline - a small town which sits right on the rail line between Columbus and Cleveland, and which used to be a big railroad town. It's interesting to think of how my hometown might change if passenger rail were to actually happen (though I'm skeptical, as versions of this plan have been proposed for decades).
That said, there are some major problems to overcome if we wanted to start doing "state-scale" transportation like that - Columbus is a very car-centric city, and in the scenario where I would move home and ride the train to Columbus when I needed to go to the office, it's very unclear how I'd get from the station to wherever I'm going.
Furthermore, there was just a very public and catastrophic train derailment in Ohio, which can't do much good for public confidence in commuter trains.
Either way, the next decade or so in Ohio will certainly be interesting. Columbus has a burgeoning tech community with lots of VC investment, and while I'm not sure where I'd put the odds, it's not impossible that our little flyover state could become the proverbial "heartland tech hub".
I still get Zillow emails from my time living in the state. $500k is in the rearview mirror. A lot of new developments in the "nicer" suburbs are hitting the $1MM mark. $500k buys you a new home in the far, far suburbs. $250k gets you a 800 sqft 40s-50s ranch in a "crestfallen" neighborhood (and the price history on Zillow will show it as being sold for $60k 10 years ago).
Many of Intel's existing fabs are in the US, and they've never really stopped building them here - particularly in Oregon and Arizona, and they continue to build there. The only difference is that now they're building on yet another site. The bulk of their fab process R&D is in the US, and they perfect the flow for a given process here and their international fabs then follow suit.
Why do they continue to build in the US? The CEO answered this over and over during his campaign for the CHIPS act: Running fabs is significantly more expensive than labor costs. They don't really save that much by building abroad.
When it comes to cheaper things like packaging, the labor costs do matter, and so they do that abroad.
So no, Ohio won't be foxconned. If Intel doesn't do well, Ohio will be dead, but they're not going to build elsewhere to save on costs. There isn't much to save.
I thought figuring this out is why the big shots get their 50,000,000 dollar paychecks :-)
Because of this approach they divested out of the memory business -- which has kept them out of the growing solid state business, screwed around with Optane, lost their modem business, have lost the GPU business at least once (who knows if they'll keep doing the ARC GPUs so maybe it'll increment up again), ARM chips, and numerous other bits and pieces over the years. Now they're losing their core business too and have nothing else to really grow and focus on.
Each of these is a multi-billion dollar market with core providers who dominate the market: Samsung, Broadcom, Nvidia, etc.
Now they're going to start selling off the bones of their business to compete with TSMC, a company with actual focus.
Once that goes away, what? Intel will contract TSMC to make their chips? I can see a financially focused MBA-type executive team seeing that as a way to eliminate expensive Fab R&D, construction and OEM liabilities -- put it on somebody else to deal with that!
Optane is an example of where Intel tried to "fake it" but ended up producing a confused product that didn't leverage common industry knowledge in either the memory or solid state storage spaces. More towards a near future, Intel should be looking at compute-on-memory hardware, which promises to radically improve computation in a number of areas, and fundamentally would give them an "in" into neural network-like tech.
Similarly, the companies that do produce NAND and SSD technologies seem to have figured out how to make money in a commoditized and highly competitive market.
Why they failed to take the market seems to be both a combination of lousy marketing (is it a drive, or are they RAM replacements?), selling only enterprise-class products, high prices, and competition technology that is both "good enough" and beats it in all of the other areas.
Spec-wise, a Samsung 2TB 990 Pro, is about as fast (sustained, IOPS, etc.), a 10th the price, and is coming from a manufacturer with a reputation of sustaining development and releasing newer, better products in that product-line. If the part fails, you're likely to not only replace it, but upgrade it for the same cost.
The before times were so awful. A sea of proprietary interfaces, dubious drivers. Intel identified a huge weak point in the ecosystem & created radical change where it was needed, taking something exotic & astronomically priced & building a stable better future, around what has always made PC so exceptional: standards. The flash industry was a mess, & Intel more than anyone made it real.
[1] https://en.wikipedia.org/wiki/Intel_Arc
"Intel Battlemage, Celestial GPUs Are Booked in at TSMC: Report"
https://www.tomshardware.com/news/intel-battlemage-celestial...
Order recently deferred to late 2024, though.
Modern Intel would have shut down the Haifa team, which would have been an existential problem when Netburst reached the end of the line unexpectedly early.
At that scale, you need to keep teams/options around simply to retain expertise and hedge your tech bets.
[0] https://en.wikipedia.org/wiki/Tremont_(microarchitecture)
[1] https://en.wikipedia.org/wiki/Gracemont_(microarchitecture)
[2] https://www.amazon.com/GMKtec-Nucbox5-Desktop-Computer-Windo...
The Atom could have been bundled as an easy to integrate SoC for things like set top boxes, routers, and smart devices running Linux/Android The sort of uses that go for MediaTek or AllWinner ARM SoCs. Intel had all the peripheral IPs to make competitive offerings.
Instead they put the Windows millstone around the Atoms' neck and the chip sucked. Intel based tablets sucked because Windows was a dog shit OS for low end devices. The few Atom based Android tablets that got made were pretty decent devices.
I've got a bunch of Bay Trail devices lying around. The ones I've put Linux on end up being very useful little devices that run very well. My Windows Bay Trail tablets are just e-waste at this point. They're difficult to use at best and typically just slow pieces of crap.
https://ark.intel.com/content/www/us/en/ark/products/codenam...
https://ark.intel.com/content/www/us/en/ark/products/codenam...
The Atom lineage also lives also lives on as the E cores in current Intel desktop and laptop CPUs. Its also coming to the server market next year as the high density, all E core Sierra Forest.
Essentially Philips had spent 30 years selling the bones of the group, leaving just a skin-suit. Which they’re now selling as well.
They've been using TSMC for various parts/chips for probably 5 years now.
Intel lost money because computing started moving towards GPU architecture and AMD has both beat them in low power CPU and offer cheap GPU.
That's all. They were too late to GPU sales.
You explained the core of the problem that usually gets little blame. Successful companies get infiltrated by Wall Street MBA types who make decisions for the short term to manipulate the stock price because their own compensation structures help encourage that behavior. Companies, especially like Intel should be built for being resilient over decades instead of maximizing profits by the quarter.
This really applies for Intel, but for Google, I think it's less so. I've been thinking you have to look at incentives and at a large corporation everyone is incentivized to make the most amount of money in the shortest amount of time. If you really want to spend years hammering away at something, it made more sense to just start your own company doing that where you would reap most of the benefits through equity. There is no reason for someone to spend years at Google on a project that is seemingly going nowhere; they would struggle to get promoted, and if it does work out, most of the value will go to existing shareholders.
Secondly, you are also at the mercy at your VPs, who got to that position by also maximizing short term results. A VC may not have the tools to force you to pivot if you are making ump-teen million dollars in year 1, but a VP may decide to kill your project to move more experienced engineers to another pet project. A passionate engineer at a pet project is an anomaly at a large corporation, who also has to fight the political winds of people trying to maximize their short term gains. And I don't think those people are wrong, it's the rational thing to do, especially when any equity upside will go largely to existing shareholders.
[0]: https://berthub.eu/articles/posts/how-tech-loses-out/
"Only the paranoid survive" applied fractally at all levels in Intel. This meant that there were very few "anointed successors" as people who had specialized knowledge regarded not sharing it as a survival technique.
This meant that as the technical people who built things got old and retired a lot of institutional knowledge simply walked out the door.
Technical excellence rests on training and sharing. Intel was the exact anathema of that.
At the time everybody was saying "of course Intel isn't going to fab AMD chips, that would be crazy", but in the end the crazy thing was giving up market dominance.
ARC GPUs is basically Intel expecting the market to be the beta tester for a very expensive & inferior product.
have you seen GPU prices lately? The intel a770 is quite a good deal, especially considering you get 16GB of vram when the competitors only offer this at a far higher price. I kind of agree with you that the product has had issues at launch, a lot of them are resolved however. Also, linux support is superb compared to nvidia.
See also Boeing (?).
Small teams, with the help of tech/AI can now move so much faster than big teams. The communication and coordination overhead is starting to be too big of a burden, especially when you can have 1 person do the work of 30-100 (using AI).
But small teams can’t buy expensive hardware up front, and modern AI is hardware constrained for many purposes. These companies will either end up renting hardware from cloud providers, or using AI APIs from cloud providers.
There are a handful fabs around the whole world who can even make those chips intel and tsmc can do.
No one will just replace any of them in any near capacity at all.
And the entry to this level of tech is not 'just a few millions' ist actually (just making a fab with people and processies of people who know how to do this) is 10-20 billions.
And if you would take any software company i also disagree: I was working in a small company, they didn't knew security really. I intriduced a ton of security. In comparision to the huge company i'm now: The developers still don't know much about security but the security teams of my company do.
In my opinion the gap will be even bigger because small companies (besides the cool and hip new elite startups) still have more of a traditional mindset while the big onces need to compete. They need to use AI and co and they have the financial background to create AI teams.
That’s quite the statement. I highly doubt that will be reality in the next 1-5yrs, especially for novel R&D work. Even more so for products in the physical world.
Their foundry business is dead...they won't compete with Samsung and TSMC. Even GF found that was a losing proposition and they had a lead on Intel and endless pockets of cash.
Steve Jobs has the best explanation for this. https://youtu.be/P4VBqTViEx4
In the 20th century, we saw the rise and fall of General Electric, Westinghouse, GM, RCA. Huge American corporations with products almost everyone interacted with almost every day. In the 1960's these were the corporations to beat. In the late 1960's, bean counter infiltration began. Of these companies, only GM isn't a hollow brand licensing shell, and it had to survive a government-maneuvered corporate bankruptcy to make it to 2009.
[0] https://www.tomshardware.com/news/intel-ice-lake-xeon-w-to-c... [1] https://www.lenovo.com/us/en/p/workstations/thinkstation-p-s... [2] https://www.dell.com/en-us/shop/workstations-isv-certified/p... [3] https://www.hp.com/us-en/workstations/z4.html [4] https://www.apple.com/mac-pro/
I'd imagine other manufacturers are in a similar boat. They put in orders for Skylake Xeons for machines that sell slowly. Pandemic sales went into laptops for home offices rather than $5k desktop workstations. Desktop workstations have never flown off the shelves but are flying off more slowly than normal.
Ryzen and i9 laptops and a beefy VPS have been good enough for many WFHers so no one is buying desktop workstations. Manufacturers aren't going to jump to new Xeon parts if they're sitting on old stock.
Even retail knew they were estimated to post a loss. I'm guessing something in the earnings report was "less bad than expected" in order for the stock to be up 5% today?
People often assume it "should" move in the direction of the overall sentiment (good earnings = up, bad earnings = down), but that is not at all true because earnings reports are a retroactive view of information people have already been trading on.
Actual results came in better than expectations (hence the "delta" you are referring to)
What metric did Intel beat on?
This quarter EPS/revenue/some metric/growth? Or guidance for upcoming quarter less bad than expected?
All it takes is one or two "good" things and everything else with "in line as bad as expected", as you know.
You don't want to go where the ball is right now. You want to go where you think it will be in the future.
I don't usually do sports analogies, but my kid recently started playing soccer and I've been trying to get her to intercept rather than chase people. Heh.
I think one of the biggest concerns right now (maybe always?) in the stock market is "things will stop being gangbusters, the economy will slow, growth will slow, and the ball with stop being projected to be in front of you/forward at any scale/pace"?
The same thing often happens in reverse: "Nvidia posts record profits" and stock tanks because an even bigger record was expected.