> I already know what the reaction to this essay will be. Half the readers will say that Microsoft is still an enormously profitable company, and that I should be more careful about drawing conclusions based on what a few people think in our insular little "Web 2.0" bubble. The other half, the younger half, will complain that this is old news.
With 17 years of hindsight to our benefit, my reaction is both of these things at the same time.
Fascinating to me that this only mentions phones once in passing, and even if it had made a big deal about that seismic change in the industry, it would still be wrong. It doesn't mention Microsoft's struggles in the console market, but would still be wrong. It mentions Linux once, doesn't talk about servers, databases, S3 or EC2, but it'd still be wrong.
Dunno what the lesson is, and obviously you can move the goalposts to make 'dead' mean something else. Maybe we'll see this again with all the companies that aren't quite AI enough today. Sure is nice having enormous amounts of cash on hand, though.
In 2007 Microsoft was on top of the console market, which was also a pretty new market for Microsoft in general. It doesn't factor into their argument much.
In any case, the article is definitely hyperbolic. The overall statement that Microsoft is no longer an IBM-tier titanic bully is correct, "dead" is an aggressive overstatement.
I think the title was (perhaps intentionally) overstated, but the content of the essay is correct. The better title (in the sense of more accurate, perhaps not for getting clicks) would have been "Microsoft Is The New IBM". And that is still true.
No, it's not correct. MSFT is a $3+ trillion market cap company doing $200+ billion revenue which is growing mostly organically at a very good clip and has a product line people and companies use mostly by choice.
IBM is 20x smaller by market cap, 4x smaller by revenue, they are shrinking organically because they forgot how to make anything and no one starting any business from scratch would use any of their products.
He got it completely wrong. It is what it is. If you comment on enough things you are going to be very wrong sometimes.
Microsoft in 2007 is where IBM was prior to that. Which, by the way, was once invoked as the prototypical example of the "old guard tech company reinvents itself". Also, the basic thesis of the essay is that new companies don't worry about Microsoft squashing them anymore, the way they once had (and did about IBM before that). That is, I believe, still true.
I don't think that is as true now. Microsoft has made in-roads in a number of markets, which has to concern some competitors. Within the last few years they left slack in the dust by bundling teams. They also bought github just a few years ago. A couple years before that linkedin, which there are still no real competitors to. Their cloud market share has been increasing and they're increasingly pushing the integration of the Microsoft 365 suite and github with azure, which has been pushing more enterprises towards their offerings. Microsoft has also started throwing money around to organizations which they want to work with, such as OpenAI and G42, effectively choosing winners.
When IBM bought/merged with Red Hat, it seemed like a big deal. For that matter, IBM was a big reason why Linux became a mainstream technology. IBM doesn't move like this any more, but then we haven't seen Microsoft's investment in OpenAI actually pay any dividends yet, either.
Nadella is extremely effective and aggressive in his moves. He always impresses me in that regard.
While Ballmer built a solid cash foundation - he would never have done a move like Linkedin or Github acquisition. And I would not be surprised if Github acquisition was a part of the plan about AI.
But I don't think it is? IBM is, in essence, irrelevant. It keeps milking legacy customers, but it has no competitive edge otherwise.
Microsoft was never this far gone. PG was proclaiming their death because startups no longer feared them. But they no longer fear Google either, not the way they used to. Google's last truly disruptive and unexpected projects - Android, Chrome, Maps, Docs, etc - happened 15+ years ago. Now, the company sticks to established niches instead of trying to constantly reinvent themselves (or when they try, they almost always fail).
There's a spectrum between "dead" and "taking on the world".
I mean, yes? IBM is profitable and is spending some of its money on projects we all benefit from - just like Microsoft, Google, Oracle, Apple, and just about any other larger tech company out there. That doesn't mean that their core business is on a healthy trajectory. Transmeta was paying Linus' salary for a good while, it didn't save them.
People blame Steve Ballmer but I think he did great stuff. He built the huge enterprise salesforce and got Microsoft a leg in the enterprise, where margins are much higher and consumers much less fickle. He also chose a very good successor and removed all the old timers who would have made trouble for Nadella (Muglia, Sinofsky to name a few). Microsoft went from mostly an OS+ consumer software company to this enterprise gorilla which will now have an IBMesque lifetime (meaning will survive for very very long) regardless of whether their projects are very good or not.
Sure he couldn't beat the iPhone, but the iPhone is the most successful consumer electronics product of all time.
Ballmer is a good used car salesman. He sold what they had, but new products suffered under his leadership. Just like Apple stopped inventing after Jobs.
Nadella built new product categories for the company.
I think two more years with Ballmer would have made the the prophecy valid.
What new product categories did Nadella built that Ballmer didn't start?. Azure was started and then prioritized under Ballmer. Office for iOS was built under Ballmer but not announced to give Nadella a quick win on his start. Ballmer set his replacement very well on so many levels.
> Ballmer is a good used car salesman.
Ballmer was a very very good salesman. The reason Azure or all these new shiny MSFT products (e.g. teams) have a leg is because Ballmer already sold AD to almost every corp out there. Microsoft went from being 0 in enterprise to bigger than anyone else and that is now the main money printer and also the future of the company.
One thing everyone forgets is Nadella was nowhere to be found few years before becoming CEO. Ballmer made a very very good successor choice elevating Nadella quickly in few years.
> WSL, becoming again a Java vendor, adoption of Rust, FOSS culture on DevDiv, VSCode eating VS licenses...
These are all nice things for developers, but I don't think they are actually good for Microsoft in the long-term or their bottom line in the short term. At least not the way they've been executed.
If anything, the treatment of Microsoft's older cash cows under Nadella has been disastrous, essentially throwing goodwill and money away - e.g., the quality of and feature development in Windows.
Were you involved with the project? I worked at msft back then and distinctly remember the office ipad team. I think ballmer actually let them release just the OneNote standalone app in 2013 iirc.
I’ll never understand this narrative. What product category has Apple missed or failed to execute well in since Jobs died? Do you think there’s an iPhone level hardware product out there that Apple would be shipping today if only Steve were still around?
Off the top of my head, since Jobs death Apple has launched AirPods, Apple Watch, Vision Pro, Car Play, Home Play, M-series MacBooks, and they’ve greatly expanded their services business.
Windows XP thru 10, Xbox, Windows Phone and Azure are just a few of the things Ballmer oversaw. I am definitely missing a lot of things. But remember he over saw Microsoft as CEO for 14 years.
He may look and sound like a used car salesman but was anything but.
Yes, Nadella did end up removing Windows Phone but you had to give it to Ballmer for standing behind that thing even when it was, at times, very silly!
As an aside, that was something I think that is one idea Ballmer did ok on. The focus on developers as an admittance that Microsoft cannot do everything and that you need to help others work with them. Developers (x 3), may have been silly but it sure is a good stance to get behind.
By contrast, as much as Apple talks about the App store, they seem to be becoming more inward focused on a lot of things. For better or worse, we are yet to know.
Anecdotally I’ve found that I often can’t get started on a new piece of work because for every possible solution I imagine all the possible ways it could go wrong. After a glass of wine I lose my hang-ups and I can “just get started” and bang out some code.
I suspect it’s the same as the well-known “social lubricant” effect of alcohol, where it eases anxiety and uninhibits nervous people.
However, while I’ve noticed it gets me going it has the same deleterious effects as in social contexts: the code I end up writing is not as clever as I thought it was the night before.
No matter, had sex! I mean… got an MVP working that can be nurtured into a relationship… I mean product.
This blog post is actually a good indicator of how ephemeral web applications are. The blog post contains a reference to some "Snipshot" webapp. But the domain is dead now, and it's impossible to get the software. It's even hard to see how the software ever looked like.
In contrast, we can all install Lotus 1-2-3 on today's machines, via emulation.
Writing webapps is writing something that will never exist in the future. At some point in time the server will be turned off, and nobody will see your work.
what I think most people will miss is the Microsoft from the 90s is very dead.
What they have now - office 365, linkedin and azure + a dose of minecraft and Marina Linux are nothing like those days.
Days when to succeed in software you had to be 10x better than microsft or they would release a half arsed competitor and destroy your business just because they were microsoft.
PG wasn't wrong, it's just that Microsoft was resurrected later in 2014 by Satya Nadella.
When Paul wrote this, the industry was witnessing a technological revolution across the board (smartphones, AWS, Google docs, etc) while Microsoft was drugged and lacking behind on all fronts. At the time, their latest product launch was probably Windows Vista.
It was a different company back then and it's such an amazing feat of Satya to manage to steer a ship of Microsoft's size.
Microsoft is still here and still Microsofting. They've had a few setbacks to total dominance. They fumbled the mobile ball and found that they couldn't compete against Linux and open source, which enable new forms of computing at the largest (cloud) and smallest (embedded/Raspberry Pi) scales that don't fit with their business model.
So they successfully pivot to co-opting the dev stack used in the open source world. That's their play with Azure, WSL, Visual Studio Code, and the GitHub and npm acquisitions: to make it impossible to do webdev on an open source stack without MSFT involvement, and hey, why not try out these other great MSFT products and services while you're at it?
It's a clever play. But they're not dead, irrelevant, or your friend.
It's really hard for a company to die once they reach a certain size because they have the resources to pivot, even if the pivot seems a bit late.
The two most extreme examples of this are RIM (now BlackBerry) of the mid-2010s when it was clear the smartphone market was no longer viable for them, and Apple of the 1990s before the return of Steve Jobs and acquisition of NeXT. When asked what he'd do with Apple in 1997, Michael Dell famously said "I'd shut it down and give the money back to the shareholders."
But both companies are still alive today (Apple more than BlackBerry) because they were large enough before they had to pivot.
Other big tech companies like Microsoft were never close to death at any time. Sure, they had periods of stagnation where they had to pivot (as mentioned in this blog post), but it was more like a cold than a life threatening illness.
Microsoft is dead indeed. Windows went straight downwards since Windows 7. All their products are not good. Can't remember a single MS software or service I'd use. They bought some companies like GitHub which are doing good, but that's not MS and they couldn't even buy OpenAI.
About two weeks ago, I had an epiphany that ‘Apple was dead’, since any cheap crappy device could act as a terminal to an AI infrastructure delivering a perfected and idiosyncratic user interface. Now, I’ve reconsidered.
This is classic demonstration of the fact, you do not have to get most of the things right, but extract maximum value when you are right.
We have the excellent advantage of hindsight but definitely in 2007 the general agreement was Microsoft was losing to Google and Apple, which were the darlings of the stock market while Microsoft remained stagnant.
> definitely in 2007 the general agreement was Microsoft was losing to Google and Apple
In 2007, [1]
MS had market cap 274B (#4), revenue rank 49.
Apple had market cap 80B(#33), revenue rank 121.
Google had market cap 143B (#17), revenue rank 241.
MS had, and still has, many billion dollar sub-businesses. Google and Apple have never had as many highly profitable lines as MS does (and did).
I think only pop media sold the story that Microsoft was "losing," and maybe only that in areas MS didn't historically compete in or make much money in. Google had almost all revenue from ads (and still do, a very one trick pony), and Apple had a few lines (iPhones, iTunes, not much else), while MS had Client, Server and Tools, Online Services, Business Division, Entertainment and Devices, each with many billions in revenue (and many of them containing yet more billion+ subdivisions). Even Visual Studio had over a billion in revenue annually.
Apple and Google had very little effect on any of these revenue streams in 2007. At best, Microsoft was late to the areas Apple and Google were in, yet took markets anyways in some areas (Azure vs google?).
just like yc likes and other accelerators like to bet on founders - they too should look at company structures / incentives etc to evaluate the future of certain companies.
from afar you can tell microsoft environment breeds killers, unlike google and other tech companies besides facebook i.e zuck & oracle.
which is why is you can never count microsoft n the likes of intel or even in terms of other industries ford out.
look at how microsoft came and knocked down the tools in terms of developer tools. look at slack. the rest of the world uses teams - it comes free.
look at how they did in the cloud. look at so called 'AI' satya or not -- microsoft has the industry by the balls and they know it. fear of gvt is what keeps them at bay.
why people are assuming this essay main points are wrong?
it's very correct.
Microsoft windows and office only make money with kickbacks to govt buyers. see the documentary on Microsoft and EU. lots of money is moving under tables.
azure is a joke that mostly gets sales based on price for people who need to buy a cloud just because. in this regard they are a walking corpse like ibm was in the 90s.
then the money came from xbox, which they tried to kill several times and only got market share because they sold at loss early just because that's what they do without even thinking. Salesforce, and LinkedIn, which they bought after they were monopolies already (and owners only sold exactly because of the fear of saying no, which the article paints very well) and couldn't run to the ground no matter how bad and abusive it got. and lately openai related stock memes despite being the least protected player (as in no hand in chip design).
what exactly paints the article as wrong? they just got a few luck hits on the way down.
By that metric Uber and Tesla are the better companies on the history of the universe.
Wake me up when that speculation translates to profit that is not just fraud of funneling investor money from AI companies they are on the board back into azure profit center.
> Salesforce, and LinkedIn, which they bought after they were monopolies already (and owners only sold exactly because of the fear of saying no, which the article paints very well)
you are right. i don't even know which one i was thinking when my finger decided to type salesforce. well. I only know enough about MS for the occasional checks to keep it into the not buy pile.
Why bring personality into this? Paul Graham made certain observations, conclusions and predictions about a very successful company - to the point or not. That is not evidence of a causative agent in terms of a massive ego.
It's almost as if we shouldn't treat their words as gospel. By them I mean people who were in the right place at the right time and had the right amount of luck, which is what success looks like usually.
You’re amazed that someone wrote a bunch of articles about opinions on the future of technology and they didn’t all turn out to be true? An expert’s opinion being wrong one time invalidates the expertise?
I don’t know. I just had to buy, and use, a Windows 11 pc. Remarkably, it shares a desk with a new M3 iMac, side by side. And equally remarkably, they’re set up so they behave almost identically, except that Windows 11 is a little snappier and has a few fewer keystrokes than the Finder for most operations. (And search works!). I quickly acculturated to it after a decade away. Entirely unexpected.
Are you noticing ads within Windows? I haven't used windows in over a decade and hear there are a lot of ads you see throughout the OS, which is something you don't get on macOS.
I use Windows 11 Professional at work, and I haven't seen a single ad.
There might be sponsored apps installed or something, but I don't care. I open the start menu, search for the application I want and run that. They're not getting in my way, if they are there.
That said, I prefer late-Windows 10 to current-Windows 11. But I hated Windows 10 when it launched and they fixed the key annoyances, so there's a chance they'll do it with Windows 11 as well. They've already fixed some, like the centered task bar.
This is false; there are ads in quite a few places in macOS, examples of which were shared in another reply to your comment.
While I'm not a Windows user much anymore, my understanding is that there are at least unofficial hacks/tweaks you can apply to remove many of the ads in Windows. In contrast, certain ads in macOS (e.g. the one nagging you to upgrade your iCloud plan when you're near your storage limit) cannot be disabled through any mechanism that I've been able to find.
I pretty much stopped using MacOS because they put so many fucking ads in the OS by default and smugly expected you to just accept that they were there. It's not as bad as Windows yet, but honestly it's just as annoying.
I'm just one data point, but I haven't personally seen any of those ads in macOS but I don't use Apple News. Also I see all of these links are reports from 3 - 8 years ago so perhaps some of them don't occur any more?
Regardless, it's rather rude to call people names here when we're just having a discussion. Your tone is offensive.
You were wrong. MacOS has advertisements, that's the long-and-short of it.
I correct people when they make this mistake because Hacker News has a habit of giving corporations undue lip-service. You can be mad at me if it makes you feel better, but the problem is that Apple doesn't respect your attention. MacOS today literally violates it's own Human Interface Guidelines to serve you advertisements; think on that for a moment.
It appears you have missed the point. I'm not upset with you for referencing reports; my frustration lies with your condescending tone and name-calling, which contribute nothing to the discussion. Your approach is frankly very off-putting.
I don’t even know what people are referring to exactly when they say Windows is full of ads. Windows 11 is my main OS, it’s the basic Home edition.
The only ads I see are when I mistype in the start menu, and get some web search results including ads, just like when I google something.
Inline web search results is a crappy feature to me, but it didn’t bother me enough to go regedit mode and disable it. It only appears when the computer search can’t find anything, it kinds of fall back on web search.
It seems the most outraged people are those who never use Windows and rage at the fact that it’s still a popular desktop OS.
Or maybe I’m just lucky and I don’t get the horrible in your face ads everywhere, for some reason.
What bothers me is not that it displays web results but that it forces Bing results and forces Edge to open said links despite it not being the default system wide web browser.
Yeah it’s even embarrassing for Microsoft, having to lure people into using edge. That may be the raison d’être of these inline web search results.
It’s a crappy feature anyway, I never use it. It doesn’t get in my way so it’s not a big deal to me. In any case if that’s all there is to this ad thing, the drama around it is surprising to me.
In what way? IBM wasn't prominent in as many areas 20 years ago as Microsoft is today. Today Microsoft has the leading business social media platform with linkedin, owns a leading consumer git platform with github, has a growing share of the enterprise cloud market with azure, now owns several of the most recognized game publishers with their acquisition of Activision and Bethesda, develops two of the top 10 most used programming languages with TypeScript and C#, and still has the leading os.
Windows is in cash-cow mode. It's a mature product that dominates a slowly shrinking market. And in terms of revenue it's not even that great compared to everything else they do. There is no point investing a lot of money.
That money goes elsewhere: pivoting MS Office to the cloud and pushing MS Azure - both by improving Azure and by improving developer tools like VSCode or languages like C# and the rest of the .Net ecosystem. And on the side they have minor projects like Github and LinkedIn.
> "So if they wanted to be a contender again, this is how they could do it:
Buy all the good "Web 2.0" startups. They could get substantially all of them for less than they'd have to pay for Facebook."
"And I happen to have some Web 2.0 startups for sale, what a coincidence!"
It’s ok to be wrong. I’m having wrong insights all the time. You just have to acknowledge it and move along. As long as you don’t try to make it look like you were actually right given some amount of mental gymnastic.
I completely agree. I was trying to say it’s okay to be wrong if you are able to acknowledge it instead of doing mental gymnastics to justify it. I don’t know where Paul Graham stands in this regard.
Paul Graham has been getting a lot of criticism recently, both from this obviously bad take and because of his praise for the fraud-ridden Lambda School and Paul's hero worship of Austen.
I personally appreciate the reality check that not everything these famous people think is worth writing down nor sharing.
This was not a bad take at all in 2007. In fact, it's really incredible just how much things have turned around, with Microsoft (via OpenAI) currently far ahead of Google.
Many of us Hackernews gushed over Elizabeth Holmes and Theranos back when they were at peak hypecycle. The thing about these fraudsters is they are good at bamboozling even smart people, especially smart people who are not in the scammer's putative line of work.
This is like saying your evaluation of my property, back when it was in bad shape was wrong... because after renovating the entire thing and fixing its issues, I got a different price.
From the linked article:
> Is there some way Microsoft could come back? In principle, yes
So, of course if you manage to change the input, you'll get a different output.
Today's Microsoft (WSL, Edge, Bing, Azure, VSCode, copilot, etc) is so different from 2007 you wouldn't even recognize it as the same company. Just like what Paul Graham has said they needed to change if they were to survive.
Turned out to be fraudulent. Didn’t provide a meaningful education, lied about placement numbers, lied about their financial incentives aligning with the students'—the list went on.
The CFPB, a federal regulatory body, came down on them (and the founder in particular) a few weeks ago [0]
in 2019 - 743 folks are on the hook for the ISA - 516 graduated - 132 got a job
I'm not sure what to make of this. What is ISA? Isn't this like 20% hire rate from taking Lambda course?
Or was it that they attracted a lot of people who simply can't make that leap from student to developer because they never had the aptitude for that job?
Not saying Lambda School wasn't a fraud but seems like they started out with some good intention but halfway got greedy and dishonest because Paul likes "edgy founders"
Just a ton of broken dreams, financial ruin not unlike some of YC's backed companies.
The way Lambda School worked was, you didn't pay for the instruction up front. Rather, when you got a job, Lambda School was entitled to take a chunk of your income.
It had the whiff of scam to me, as it reminded me of the payment arrangements for another outfit, variously called Unbounded Solutions, BrighterBrain, and EnhanceIT: you go to Atlanta for two months of "free" training, after which you were bound to indentured servitude to them for a period of two years. Attempt to break the contract and you would be billed a minimum of $20,000 for the "free" training they offered you.
What Paul wrote in 2007 could be true, while what you wrote about Microsoft in 2024 could also be true.
Microsoft changed. And for the better. The death was short lived, but in many ways, especially in the cultural zeitgeist of software engineers and tech folk, Microsoft did die there for a bit. Satya did a great job of resurrecting it. He has been masterful in his role (I worked at a company that was acquired by Microsoft).
Companies who were once the 900 pound gorilla can die, despite making a lot of money. Oracle still prints money, but how many people under 50 use Oracle on a day to day basis when they are building a company? I would venture to say approaching zero. Google is still printing money, but in many ways it is dead (or dying).
I just don't believe this. It's like saying that Apple "did die there for a bit" when they were selling shitty Intel Macs; it's not true. They were still alive, well, and had a significant market share of incredibly enthusiastic customers.
Microsoft had been a Department of Defense contractor for more than 20 years at the time PG published this. His conclusion was narrow-minded then, and it's narrow-minded today.
> but we've never bothered to invite Microsoft. Nor has anyone there ever even sent us an email. They're in a different world
> I'm now surprised when I come across a computer running Windows.
Funny how PG didn't have enough perspective to realize that he himself could be the one in a different world. A bright guy, but like you say - narrow-minded.
As a developer, I have a really poor view of companies that choose to go with Microsoft technologies (other than typescript)
It’s purely a cultural thing, but from where I’ve worked and what I’ve seen, companies who choose Microsoft stacks and services (azure, C#, windows server, etc) are very management heavy and most are hostile towards their engineers, whom they see as cost centers.
I think that stigma still clings to Microsoft today
(1) rolled out unstoppable, inscrutable "telemetry" in Windows,
(2) decided that "opt-out" was appropriate for privacy-invading features, and
(3) decided it was okay to silently reset such opt-outs.
(4) most recently, decided it's okay to violate copyright for any material they can scrape off the web, except for victims who can afford to fight Microsoft in court.
I don't know any of their actual reasons for those policies, but as an end-user their reasons are indistinguishable from "because we can".
IIRC, in the MS-DOS / Windows 2 up through ~early Windows 10 days, Microsoft's sleazy behavior was mostly limited to their interactions with other software companies.
> "Oracle still prints money, but how many people under 50 use Oracle on a day to day basis when they are building a company? I would venture to say approaching zero."
Oracle is having something of a resurrection moment lately, with Oracle Cloud becoming a surprisingly popular option for startups.
"At least six venture-backed AI startups, including Character.ai and Adept AI Labs, primarily rely on Oracle for cloud computing, according to the companies. Two founders of AI startups who have used various cloud providers told The Information that Oracle can run complex machine-learning models more economically than Amazon Web Services or Google Cloud."
In a sense Nadella’s Microsoft also bribed their way into popularity with startups, with both free cloud services as well as free software like VS Code.
As someone with no skin in the game and also someone who thinks that MS, Google, Apple, etc are not really “better” from an ethical standpoint than each other/any other big tech company: I always get the impression that Satya is a brilliant CEO and on the other hand Sundar is really not that good. Is this the general consensus?
I don’t think he did. They are not the innovative company they were under Satya. Sure the stock price is up by what are they actually doing? Office isn’t the force it was, teams and azure are just copying others and they utterly enshitified Windows. Xbox has also gone backwards.
The antitrust issues were a small piece of a much larger puzzle. Microsoft wasn't dead because of anti-trust issues. They were dead because Ballmer took them in the wrong direction. He made lots of money, but killed the culture of Microsoft.
If they wouldn't have had the eye of regulators on them they could have financially buried a bunch of startups. But that didn't make them turn their products into garbage. There were many issues.
Back in 2007 to early 2010s? Sure. Certainly not these days though they more or less reinvented themselves (e.g. Windows is basically a side gig at this point, much like Android is for Google).
Well neither IBM is doing much in terms of business machines lately. They are still a second tier giant with enormous turnover. They own a few pieces of popular tech either via acquisitions or licensing agreements but none that came from within the business. Nobody in the industry particularly cares what they're up to.
With 17 years of hindsight to our benefit, my reaction is both of these things at the same time.