Microsoft, currently the most valuable company, is having a Nadellaissance
bloomberg.com
bloomberg.com
Instead, “Intellisense with AI now available on VS Code” is Microsoft’s nice way of saying “We harvest the shit out of your data”. Ah, like a true Ad company. Don’t get me started on Windows 10 telemetry.
I really believe that Apple’s bet on privacy will pay off well in the long term. They’re different companies (Apple is selling hardware) but boy they could have gone into Ads with a MASSIVE user base but they chose to limit Ads in the App Store. It would have been huge expansion opportunity but they didn’t pursue it. And it will pay off as privacy awareness spreads and Microsoft could join them as operating system providers.
What? The IntelliCode description says that "Contextual recommendations are based on practices developed in thousands of high quality, open-source projects on GitHub each with high star ratings.". I'm not clear on how you think Microsoft will harvest data from doing this - much less how it is related to ads.
I was trying was to make a tongue in cheek point, rather than specifically digging into exact methods of how they collect data. You may be right but it is besides the core issue we are discussing here.
At the end, it is irrelevant how they're collecting data - all you need to know is this: https://about.ads.microsoft.com/en-us
Disclosure: I work at Azure but not on VS Code
I totally understand, though about trust. We've been asking users for feedback for 20+ years, I hope we've shown ourselves to be good actors!
Disclosure: I work at Azure on machine learning
Not sure about the Azure team (since my experience with Azure is limited to a one-off project that required a SQL Server instance to test compatibility with some client's legacy DB, plus some very recent dabbling in VS Code and Azure Data Studio), but whichever team runs Windows nowadays, for example, hasn't exactly "shown [them]selves to be good actors". Or is preinstalling Candy Crush even on "Professional" versions of Windows 10 considered being a "good actor" by today's standards?
(Sure, even the first versions of Windows had games preinstalled, but those were first-party instead of third-party, they ostensibly existed to help users learn how to use a mouse, and they didn't plaster themselves front-and-center on my Start menu like seedy massage parlors in a Bangkok red light district)
Needless to say, "we've been siphoning your data for multiple decades, so you can trust us to keep doing so even though we're actively betraying that trust at this very moment" is cold comfort at best. I would hope that the New™ and Improved™ Microsoft™ would at least have some self-awareness about that.
I'm sure your intentions are good and pure. I hope they stay that way. I have no way of knowing whether or not they will, and historical precedent says they probably won't. It shouldn't be surprising that I and plenty of other tech-savvy users would take such claims of trustworthiness with a baseball-sized grain of salt.
* a lack of trust in corporations in today's economic and political climate and
* the skyrocketing growth of the online advertising business, which has made user data quite literally the gold mines of the digital age
The incentives today are very very well aligned for a corporation to sell user data. Even if a company may mostly be made of good people, without explicitly written contracts or declarations of any kind, a Company can change internal policies rather quickly. Maybe there will be a manager who sees the selling of that data as a way to boost profit and rise within the company. Maybe the company sells that division along with their data.
I don't believe you have, sorry (Microsoft I mean, not you specifically). I get the impression Microsoft is the same company it always was, only now it's better at convincing the gullible that it is on their side.
That Windows 10 update prompt for previous Windows versions were obnoxious and downright deceitful, making the 20+ years old traditional 'close window' X in the corner of the window actually download the update instead of doing what users expect it to do.
That's before we even get onto the subject of Windows 10's draconian privacy policies and use of dark patterns all over the OS designed to get people to disable their privacy settings.
Of course, even if they don't fall for tricks like coloring the 'privacy customisation options' link during installation a slightly different colour of blue than the background of the window to make it deliberately hard to read, they will still likely have all their privacy options reset and have previously removed crapware like Candy Crush added all over again in a forced update.
I'm on a Pro version of Windows 10 and I can't even disable the telemetry and constant phoning home every time I launch an application and or encounter a problem (that I absolutely do not need help with) with something I'm running.
I could understand this crap with the Home version. Not on Pro which I fully expected to have a degree of control over like I did with Win7 pro.
As a result of all this, Microsoft finally made my shit list after decades of shady behaviour and anti-competitive business practices and I won't be buying or using anything by them going forward.
Hell I even started PC gaming on Linux, that's how much I dislike the things Microsoft does.
https://docs.microsoft.com/en-us/visualstudio/intellicode/in... - No user-defined code is sent to Microsoft, but we do collect information about your use of the IntelliCode results.
For base model suggestions, which are open source or .NET types and members, we capture whether you selected an IntelliCode suggestion and log the name of the suggestion. Microsoft uses the data to monitor the quality of the base model. For custom models, we capture whether you selected an IntelliCode suggestion but do not log the names of your user-defined types or methods.
Thanks Mark Wilson-Thomas Program Manager Visual Studio IntelliCode
This assumes, of course, that y'all aren't doing any weird code-injecting funny business when packaging it up for installation :)
What most viewers saw during the Facebook congressional hearing was woefully out of touch, forgive me for using this phrase “old white guys” questioning Mark with many facepalm-able moments. But the house hearing, which includes 40 representatives from the Millennial generation, who grew up on tech, asked much better questions. It just got swallowed up by the news cycle because the Manafort Raid was happening at the same time as hearing about the Cambridge Analytica scandal was occuring.
All of this is still new. Banking, Finance, Real Estate, those are all well regulated at this point and arguably much more sophisticated or complex. The situation we have now with the Googles, Amazons, Microsoft, Facebooks of the world packaging up our metadata, combining it with other third party data, training AI, etc is all new, and those who are knowledgeable about it are just now making it into congress, so I am optimistic we will start to see real regulation on these shameful business practices soon.
Roger McNamee had an excellent conversation on the Sam Harris podcast recently about all of this. It was a very good conversation and one I highly recommend listening to.
If there are any changes to this doc - https://docs.microsoft.com/en-us/microsoft-365/compliance/gd... - you would be able to check that immediately, and we would never change our data collection policy without changing that document... It's the law!
Disclosure: I work at Azure on machine learning but not VS Code
However, if you don't trust us (and I get it!) Your best bet is just to opt out which you can do at any time, including during install.
Disclosure: I work at AZure on machine learning but not VS Code
For my own general purpose computing, I want Windows desktop especially with its backwards compatibility.
For business, I'm doing a startup which for the users is a Web site, and I wrote the code on XP and will run it on 7 as my server until I switch over to Windows Server.
I've done nothing with Unix or Linux.
In the past 10 years, I've typed in about 400,000 lines of text as software, with about 100,000 program language statements. For the compiled code, it's essentially all Visual Basic .NET with ASP.NET for Web pages and ADO.NET for getting to SQL Server database. I like VB.NET. I did my startup Web pages with ASP.NET -- seems fine to me although they are my first Web pages and I;m no Web page expert. E.g., I wrote no JavaScript although ASP.NET writes a little for me which is optional.
For a cloud, for now I'd be concerned about cost, startup time, and security. Later I'll be concerned about security.
I tried Visual Studio for a few minutes and gave up on it. I've used no integrated development environment and have no desire to.
My two most important tools are my favorite text editor KEdit and my favorite scripting language Rexx. I type my code into KEdit and have a lot of KEdit macros to help me with the code. The code for my Web site is about 100,000 lines of typing and about 24,000 programming language statements -- I had no trouble debugging and never wanted anything like Visual Studio.
I wouldn't use Visual Studio for free -- too much botheration for too little need.
When I looked at Visual Studio, I could find no reasonable, usable, competent documentation, and no way do I want to take out months to figure out like a puzzle and document Visual Studio for my own use. E.g., Microsoft keeps talking about "intelisense" as if I should already know what that meant and would like it. Of course, I can't look up intellisense in a dictionary, and absolutely, positively, with feet locked four feet down in reinforced concrete will Microsoft refuse to document, describe, and explain what they mean by intellisense. It is as if their gibberish not in any dictionary has self-evident meaning and value -- it has neither. Grotesque, outrageous, inarticulate, incompetent, sick-o communications.
It turns out, ASP.NET is super easy to debug -- just give the .ASPX file to a Web browser and let ASP code do its things.
For Office and e-mail, I use my legal copy of Office 2003 with Outlook. Now that I have good, extensive notes on how to adjust the settings and options on Outlook, I can do the setups in an hour instead of the several days as before; it's fine. There are some improvements I could think of, but I doubt that Microsoft would be interested. It might be that I could program the improvements with the old VBA which might be able to read and parse an Outlook PST file, but so far I've never tried VBA and when I did want to try it didn't have a copy. For Excel, I use it to draw simple graphs and otherwise regard it as worthless -- I'd much rather write code in Rexx, VB.NET, Fortran, PL/I, etc. and then use Excel to draw the plots. My understanding is that now there is much better graph drawing software, likely with API's, or better API's, than Excel.
For high quality word whacking, I use D. Knuth's original TeX and love it. I hate Word -- used it some, got okay with it, but hate it.
Lessons: To me, for my personal computing and for business, I want the full power of a good desktop computer. I place high value on backward compatibility, e.g., back to Office 2003, an old Watcom Fortran, an old IBM OSL (Optimization Subroutine Library) to be called from Watcom Fortran, KEdit, Open Object Rexx, etc. For business and software development, VB.NET is fine.
Far and away my greatest gripe with computing, the computer industry, and Microsoft is documentation -- on average, the quality of the documentation is awful. To me, the biggest problem in my startup, by far, is poor documentation. Broadly the poor documentation has commonly taken me 100-200 hours to do things I should have been able to do in one hour. So, I DO write my own notes and then AM able to do the stuff in one hour. The bad documentation is close to killing my startup. Long since I should have been sending six figures a year to Microsoft for licenses on their software, and the main reason I'm not is their documentation. E.g., it took me two weeks of full time mud wrestling JUST to find a connection string that worked with SQL Server -- should have taken 10 minutes. Recently I spent 80 hours full time getting simple file sharing, as a first time user, between my 7 and 10 systems. I wrote up notes for myself that will solve the problem for any first time user in less than an hour. I posted the notes on TechNet. Responses from others were awful, e.g., kept talking about Workgroups and Homegroups and some Windows password tool, ALL of which are just irrelevant down to next to useless. Apparently nearly no one still actually knows how to use the simple, well designed, command line NET commands to set up first time user file sharing. The old NET documentation is also awful, gets a grade of flat F in just simple Bachus-Nauer syntax notation 101 and totally omits anything about semantics, meaning, usage, understanding, security, consequences, timeouts, etc. I had so much trouble with even simple things with SQL Server that eventually I got help, really simple answers, from some high up SQL Server executive.
To me, the most serious problem blocking information technology, computing, Microsoft, and my work is BAD DOCUMENTATION.
When Microsoft learns how to describe their work, then I'll start to consider if they are a competent, functioning company with a bright future.
They don't need your telemetry for that
Wait, what? Is there a way to opt out of that in GitHub? I don't want their AI to scan my code.
Realistically, since most of my GH code is Common Lisp with few stars, I doubt it's looking anyway, but I'd like to make sure.
I'm sure it's covered in their TOS, but as a (still) paying GitHub customer I don't want them to do that with my repos.
{this} can be code or text or an image or any other content.
I didn't realize I wanted this, but now I do.
It is a sensitive branding issue.
SaaS companies already make ridiculous margins. Applying a sensible regulatory framework around privacy and data usage auditing would add overhead to be sure, but I’m also sure software margins will cover it.
To be fair, this has always been a problem — I think the recent failings of the historical trust model are due to better information rather than any increase in exploitative behavior. But prior to the smartphone era, trust was a fuzzy, emotional problem where a customer’s trust in a brand could be influenced by marketing alone. Today it is an explicit, quantifiable problem where users are willing to vote with their feet.
What companies are we bestowing authority upon? Do those companies’ business models truly have our best interests at heart? What guarantees do users have that their trust won’t be violated? I firmly believe that whoever can solve these questions to the market’s satisfaction will own the future.
Any good product organization (and Apple is one of the best) focuses on their customer. We all know that. But Apple goes further and takes their customers’ side — the iTunes Music Store was an absolute disaster for the recording industry but a huge boon for the consumer (and, of course, iPod sales).
Apple is only in a position to do this effectively because they have such a large and loyal customer base. They have a large and loyal customer base because they have consistently driven innovation in personal computing while making high-quality devices that people are willing to pay a premium for. Why would they betray that loyalty?
Apple’s business model is simple and easy to understand. I look at their balance sheet and see where the money comes from: hardware sales to consumers. You can do the same with Amazon and Google too — which shows you why their definition of “customer focus” is so different from Apple’s.
Out of many large corporations the only company I trust is Apple. I’m happy to pay more. I’m happy to sit through their repair bullshit. I trust their motives.
Amazon is currently expanding into markets at such a rate everyone is afraid their market will be next. Microsoft currently is staying in their lane. If someone is going to be invading my privacy I'd rather it not be a direct competitor.
Netflix is AWS’s largest customer, even though it competes with Amazon Prime Video.
Their most valuable data: what is being watched, where, by whom, is hidden from Amazon.
But they still with with Ads. And they've tried Ads in the past, with iAds. That they failed doesn't mean they didn't try, and doesn't mean they won't try again. Apple is all about advertising. Now, we can discuss how they handle privacy related to ads, but let's not pretend Apple is in no way a contributor to ad ecosystem.
Windows (likely) always could offer better privacy for a fee. You can buy a Windows computer incl. services at a price point you can only dream of Apple hardware.
Facebook/Inst/Wapp and Google show that the majority of people go with 'Geiz ist geil' (cheap is cool), privacy be damned. Awareness doesn't necessarily lead to actions (read: payments).
IMO what I paid for my Windows Pro license is already that fee and then some.
1) Microsoft had telemetry in Windows Phone.
2) The security/privacy mindset hasn't died alongside Windows Phone. I think it's actually gotten stronger, GDPR has been the biggest forcing function.
[disclosure: MS employee, but my employer wants me to state that this is my opinion, not that of Microsoft]
This mindset is toxic and needs to die.
With the absence of telemetry, how do you know Linux is doing just fine? (specifically in the desktop world, not server world).
Microsoft has an interest in knowing how well Windows is doing, and Windows has lived in a world both with and without telemetry. It was way, way harder to gauge success prior to telemetry, and it was also much more frustrating for the user to communicate their problems. I don't think Windows could have stayed in the pre-telemetry engineering world and maintained its position as a competitive desktop OS.
I ask this mostly out of ignorance, I've got experience using Linux, but I really don't know what the dev-loop is like for active engineers.
There's nothing wrong with transparent, privacy-respecting, and configurable telemetry, but what windows 10 does is none of those. Trying to describe it as something that's good for the users comes off a bit disingenuous.
You (Microsoft) shouldn't enforce people to send telemetry. Ask kindly and I suppose you'd get a better response. (VSC is a good example how it could/should be done)
I look at Linux-focused communities on Reddit, Hacker News, etc (there’s always a good rant about Linux’s shortcomings).
But that would, you know, mean you actually need to pay competent people that can understand the discourse and engage with the community, something you guys clearly don’t give a shit about (have you ever seen the Microsoft Support forums? It’s a disaster).
Ugh. I am quite proficient with tech stuff, but last month trying Ubuntu 18.04 LTS on my desktop left me pretty disappointed. A hard no-no for anyone non-technical: the graphical login did not work until I manually switched to the text terminal and installed the latest NVidia driver (Titan V).
And even now one of my screens does not wake up until I toggle it on/off.
And no freeform per-screen DPI.
As an nVidia card owner, I feel your pain. It's a case of both sides (nVidia and Linux) digging their heels and refusing to cooperate.
A bit a more friendly reply would be appreciated.. What is toxic exactly?
Linux (the Ubuntu desktop which I use) might do fine, but it needs to be financed also, developers need a salary. (Some) people donate, there seem to be 'some arrangements with e.g. Amazon' and - I suppose - the main funds currently come from Canonical using server edition income.
There's also the TVs, set top boxes, refrigerators, and cars (among many other consumer-facing devices) that run Linux (which completely displaced Windows CE).
Trying to argue "well those don't count because they're not real Linux" is 1) blatantly false and 2) a textbook case of a No True Scotsman fallacy. Linux is Linux, regardless of which userland happens to be running on it.
By your logic every developer is a Windows user, since they use Stack Overflow that runs on Windows, IIS and SQL Server.
And you bring up internet infrastructure?
https://www.theregister.co.uk/2012/12/07/stallman_on_ubuntu_...
The fact that Canonical had developed spyware has nothing to do with the Linux kernel, glibc, bash, GNOME, systemd, PulseAudio, all of which prove that you don't need ads and tracking to maintain Linux software.
By contrast most Windows users consider bugs to be something that requires an anti-bug spray.
Are there ads? I've never seen anything other than a very occasional promotional tile for a Microsoft product. Apple cross-promotes, too.
This could fuel class complaints that the poor will be snooped on more than the rich.
While they’ve made some better decisions recently, this one has shown no signs of change. It’s why I still relegate anything MS to VMs.
But Tim Cook, I would bet, would be averse to it.
Sundar, on the other hand, is just content collecting $$ and showing more ads.
Not everything is golden at MS now either. I am working on creating a platform for our medical software on Windows 10 and it really makes me want to switch to Linux or iOS. Windows is a big mess of old and new code and attempts of sandboxing like iOS but only half-hearted. Keeping up with the semi-annual channel is also insane. I would much prefer annual or biannual releases that are actually stable. It's a very developer unfriendly platform now.
Still better than trying to keep up to date on the NodeJS or frontend development world though :-P
But Microsoft's recent efforts on Linux & macOS, i.e. VSCode, has been surprisingly good. They do a lot of things, I think Typescript is run by MS as well.
So that's something that positively changed under Nadella.
Ballmer brewed the culture to be like that.
It's not like that at MSFT anymore.
In a relative way.
If you are going to talk about operating systems, iOS and Android/Linux blew by Windows on his watch despite Microsoft's big head start in the smartphone space. I think that's directly the result of Microsoft not being exciting.
If it weren't for Android, how many fewer people would be using Chrome as their daily mobile browser? How can we quantify the data slurped up and clicks generated on google.com by those added chrome instances? What about YouTube impressions? How many fewer people would be using Google maps daily? How much less location data would Google possess without living in people's pockets?
How did they measure the value of the Google play services ubiquity?
Have these things plus the app store really only generated 21 billion?
https://www.theverge.com/2016/1/21/10810834/android-generate...
If it weren't for Android, how many fewer people would be using Chrome as their daily mobile browser?
Chrome doesn’t make them any money besides Google.com being the default search engine. Pre-Chrome, Google already had an overwhelming share of the search engine market.
What about YouTube impressions?
The same with YouTube. My old Blackberry phone came with a YouTube app as did the iPhone before there was even an App Store. Would Apple have been so fast to not bundle Youtube or Google Maps if it weren’t for Android?
Have these things plus the app store really only generated 21 billion?
I would hope that Oracle’s high priced lawyers would have taken all that into account.
Considering why the number was revealed, Google had every incentive in the world to understate as much as they could get away with. Plus, how is $21 billion profit not a great thing for them? The revenue number was north of $30 billion.
Office 365 was a classic Microsoft bundled cost play -- you have this big investment in your EA around the Microsoft ecosystem and you basically pay to put email in the cloud, and get Sharepoint and Skype for free. They then made deals on EA renewal to convert "on-prem" Office licenses to subscriptions to lock you in.
In subsequent renewals, Microsoft uses Azure gift cards like a "tax credit" to drive behavior without a competitive process. They're essentially applying the Office 365 playbook to Windows Server/Active Directory/SQL Server and eventually Windows Client. Example: You pay for Azure AD by getting ATP for "free".
It's not taking away from Satya, his genius is dressing up the classic Microsoft hard-nosed deal playbook and making it friendlier in key ways, and cutting the cord on legacy. (Which is a risk, as it's an opportunity for AWS, Google, Oracle, etc to sell) But ultimately, the strategy existed and was executed before with Ballmer, but with less charisma and humanity.
Sure, Ballmer had failures too - Windows Phone and Zune didn't take off. But you can't win every time.
EDIT: Sorry, was lazy and abbreviated Windows Phone to WP.
I disagree with this too. I've been looking into .NET Core since it's been open sourced. And I'd consider Azure. But I certainly won't be using windows (for server stuff) unless I'm forced to.
It's funny that 15 years back I was so anti MS and pro Google and now I think it's the other way!
Stay skeptical of all of them. No need to be fan of one of the big techs. In the end they are all greedy organizations that want to dominate. None of them should be trusted.
Think how reliable Windows would be today if they hadn't left in all the cruft as far back as Windows 2000.
"Trust but verify" and all that.
I work in the information security assurance field, I swear that 90% of the issues I see at companies with external service providers comes back to the fact that their contract does not have anywhere enough ability to hold the service provider to task...
Get everything you need in the contract / agreement, then hope you never have to use it.
In the end you need to pick suppliers you can work with productively no matter what the correct details are.
A simpler oxymoron is "to make haste slowly". In that phrase, the "make haste" part describes velocity, while "slowly" describes acceleration. Anyone who has taken Calculus knows that velocity and acceleration can be independent variables. Thus "to make haste slowly" is an oxymoron but not self-contradictory.
Similarly, "trust but verify" describes two possibly independent variables: your attitude about the future and your behavior regarding the past. You can choose to trust people to make good decisions for the future, while simultaneously verifying they made good decisions in the past. Alternatives to "trust but verify" include "always trust", which has obvious bad consequences; "never trust", which has different but also bad consequences; "trust sometimes", which is ambiguous; and "verify always", which is unrealistic. The phrase "trust but verify" describes a healthy attitude and behavior.
Something along the lines of: "Q: Can we trust these guys to uphold their side of the deal? A: Sure, trust but verify."
Edit: here it is, it's actually a funny Russian proverb that Reagan was repeating.
The term became well known in the US during the cold war, in the context of arms treaties. The US and Russia were making agreements that relied on at least the appearance of, and some degree of actual, trust. Meanwhile they were undertaking extensive efforts to check if the other party was actually following their obligations.
It's just that "trust, but verify" is shorter and more memorable.
You should be pro-neither. Microsoft used to be against FOSS because they thought it would make them more money; Microsoft is currently friendlier to FOSS because they think it will make them more money. Does that make them bad people? No because all businesses are that way, even the vaunted "don't be evil" Google.
Yesterday I decided to write some C# code on Linux. That's easy, I thought: VSCode and .NET Core can both be found in Arch's repos these days. I just have to set an environment variable and edit VSCode's settings to disable telemetry... But alas! It turns out that I need to install Microsoft's proprietary release of Visual Studio Code because the C# debugger is only licensed to work with Visual Studio Code
Here's the Github issue about it, providing a bit of context and a lot of useless discussion: https://github.com/dotnet/core/issues/505
[1] https://github.com/dotnet/coreclr/labels/os-freebsd
The actual binary releases of Visual Studio Code
https://code.visualstudio.com/license
And here's the license for VSCode itself
"This license applies to the Visual Studio Code product. Source Code for Visual Studio Code is available at https://github.com/Microsoft/vscode under the MIT license agreement..."
Source code and final product have different licenses but VSCode and Visual Studio Code refer to the same thing. If you're using the official release channel then you should have the final product that is licensed to use the debugger.
However, placing all companies in the same bucket is committing a similar mistake.
There is definitely a “gradient of morality” amongst companies, despite all of them doing things because they think it is what will make them the most money in the long run.
A company building products that end up aiding the carrying out of genocides, for instance, is clearly on a more extreme end of the spectrum than a company making money selling music streaming subscriptions.
Couldn't that describe just about any company or open source project?
A database and office software end up helping you manage and track your genocide.
A camera company's products and photo/video editing software end up helping people create beheading videos and child porn.
When we create a general purpose product and release it to the world, we lose control of who uses it and for what purposes. In fact, that is a specific requirement for free and open source software licenses.
But perhaps you'd agree with me that there's a line between a company making products like cameras - which can certainly be used to cruel ends, but typically easily replaced by any of your competitor's products, and not pivotal to such efforts - and a company building a product that has no real replacement, is central to anti-human practices, and routinely ignores the criticisms and suggestions of human rights groups around the world.
I said that if I worked for a camera company, of course I would know that our products would be used for all sorts of different purposes beyond our control, but if the camera company was doing special partnerships with porn companies I would probably quit.
(For anyone who is thinking "what's wrong with adult porn?" I'm just sharing my feeling from that time. Substitute "child porn traffickers" or "genocidal maniacs" to get an idea of where I was coming from.)
The only explanation I found plausible, is that people like David and not Goliath. They always prefer the underdogs to win, and the titans to lose.
As for why some of us defended Google there was a long time where they - IMO - were very nice.
Big contributors to a number of good causes.
Fired sysadmin who looked into data of someone he knew, even of he had a good reason, because trust came first.
Old AdWords/AdSense with somewhat relevant text only ads wasn't too shady IMO.
... at a time when Microsoft had a CEO that said Open Source was like cancer. Etc.
Looking back my guess is it went wrong somewhere around the doubleclick acquisition.
It almost makes me want to give Windows a try again (ha not really). But I'd try a "Microsoft Linux" desktop (which no longer sounds like a ridiculous concept), before I would try Chrome OS.
IBM has free OSS labor and outsources to India. Choosing Nadella as CEO makes a lot of sense for that plan.
Satya became CEO during the beginning of 2014. And, the book was published on September 26, 2017. So, you read it when the puck was in transit.
But it was interesting to know the foundations of the transition being laid from day 1 with Nadella as CEO. The book is more like a little bit of a background of Nadella (his life in India and early days in US) and a little bit about his career in microsoft before becoming CEO and the rest of it is about the transition from pre-CEO to CEO and the initial months.
TBH, I was a bit skeptical that culture change that he focussed on would end up having such a huge impact. But I was sure that his work in terms of focus on cloud, AI, ML and also considering the challenges of such computing power, certainly was the right direction for MS to move.
Or 15 years before that, pro-MS and anti-IBM. IBM is still huge, but they've found a steady revenue stream doing enterprise stuff, so they're not creating (much) fresh evil. MS could have made steady profits writing decent desktop software (I'm not so sure now). They're both powerful but comfortable. Google is huge, and still thinks it can rule the world.
Companies are a bit like religions. When they're small, they can't have much impact. When they're big and young, they can wreak havoc. If they make it to comfortable old age, they're peaceful and stable. The Catholic Church is huge, benign, and stable now, but 1000 years ago they were a bit more conquest-focused and violent, and 2000 years ago they were just a few harmless militants.
That's a really bad analogy and detracts a lot from your point. Big religions that reach old age are merely stable, but that doesn't necessarily make them peaceful or benign. My guess is that you're probably looking at it through the lens of living in United States.
> My guess is that you're probably looking at it through the lens of living in United States.
Well, that is where I live, so that's the lens I have... My experience is that religions tend to settle down as they age. Catholics, while they may have firm belief, tend to keep it to themselves, and not to proselytize. It's more of a culture and ritual than a fervent cause. Protestants, especially Evangelicals, tend to be a bit more vocal. Mormons were killing people not so long ago[1] (though my Mormon friends never tried to convert me). Scientologists are just insane. Our older religions are, on average, more peaceful and benign.
I don't know where you live or grew up, so I can't guess at your experience with religion.
1. azure
2. o365
1. because I knew what happened (and was happening) with AWS
2. because microsoft moved to subscription model
Those two things, connected with third "secret" incredient: BIGGEST user base (corporate) which WILL use those two offerings = success.
Satya is/may be great. I bought stock around 2014/2015 because of the above and nothing more.
If you have a big userbase you can monetize it really really well.
Source: Me, at MSFT before Ballmer left and still there now.
That doesn't mean Satya's not had a good effect overall. Part of the investor postives were due to PR differences that Satya made (which are very good), and his ability to talk about the future of the company is far superior to Ballmer.
Ballmer also subjected all of those new initiatives to the Windows/Office dominance, which is why they all failed.
Nadella's most impactful, and fundamental, contribution was demolishing Windows's hold over every other department.
It moved the onus of testing on Devs who've never done it before as part of their job and thusly very lacking. The bungled W10 recent releases can be directly tied back to this model. As PMs and Devs don't account for testing like they used to before shipping to customers.
This may be fine and dandy short-term, but if you're biggest customers rely on stability for code bases over 25 years old--getting rid of focused testing and reliability is a big no-no.
Could this explain some of the QA blunders like some Windows 10 users losing files after an upgrade?
If so the move may turn out far more costly than any savings or telemetry gains.
I don't quite understand this move. Would you mind to elaborate?
Are you saying that previously SDET role exist at MSFT and today, it does not exist anymore? And that Developers have to rely on Telemetry (logging, metrics, notifications, error reporting, etc etc) ?
Perhaps I don't understand "Telemetry" in this statement.
No one really knows what’s going to happen in a recession. I suspect it won’t be pretty. Revenues will probably contract just as well in the reverse direction and I suspect many customers will look to cut costs by eliminating monthly recurring expenses.
Perhaps recurring revenue business will be more resilient. I can see a path for that as well. But no one really knows.
The difference is that recurring payments vastly improve your cashflow situation. Instead of giant pulses every 3 years you have a smoother pipeline of cash coming in all the time. Even if that pipeline shrinks, it still flows.
It's far easier to plan for a 10% reduction that happens continuously than a 3 year plan with lots of "??$??" sprinkled throughout it. And far easier to recover from mistakes.
Most companies don't optimize for cloud costs to great lengths.
Though the big corporate user base is not a certain success, IBM is a testament to this.
(Was going to say Oracle too, they haven't had a stock surge comparable to Microsoft but seem to be growing (some how?))
What is this, specifically?
Before it was extremely profitable, and it had no competition.
I would have thought open sourcing so much of it's framework, vscode, and typescript would have had a greater impact.
I have a pedantic correction here.
Instead of open sourcing the .NET Framework they wrote a new opensource framework (named .NET core) that is very similar to their closed source framework. Microsoft is currently supporting and expanding both frameworks.
It defies logic to support 2 frameworks, one of which is WAY BETTER then the other and x-platform.
The ability to build and run applications on Linux is something even old school .NET shops are excited about.
For me personally, these two open-source projects have greatly influenced my view of Microsoft for the better.
How about Apple?
Outside of Windows and Mac, you're in Open Source land, which is great if you're technically minded and want to tinker and explore, but not much of an option for folks who just want their computers to work with the consumer software and file formats around them. Other than Apple, where are the commercial alternatives to Microsoft?
I'd be much more sympathetic to Nardella's renaissance if it meant real consumer choice. The situation is better than it was under Ballmer, but not as good as it could be in a real free marketplace.
My biggest problem with Google Docs in general is not about features but that performance is pretty bad.
I can't see myself ever paying for Excel for personal use at this point.
Any Google Docs product feels incredibly sluggish, and I feel this trend lately even in Gmail or Google Keep. It just constantly feels like those old-school fat clients constantly hiccupping because of a slow server.
No issues with file formats
Amazing unbloated (or at least to the point) user experiences
I know everybody thinks they're "good enough" for most people, but they're still crap compared to the regular desktop programs.
But then again, if the stock market were rational, Lyft would have never been able to go public and Tesla would basically be worthless.
Shareholders tend to look forward rather than backward and it was getting very clear that windows and office were increasingly less critical to users and that there were some serious revenue issues around that going forward. Windows on the server was basically getting less relevant by the day. Windows desktop was feeling a lot of pressure from Apple and increasingly Google. Office sales were suffering when companies started using alternative office tools from e.g. Google. Windows Phone was not working that great either and struggled to gain market share on Apple and Google. The Nokia acquisition that Balmer pushed through was sort of the nail in the coffin here. That was an expensive failure and that is 100% on Balmer.
The keyword in those strategies was windows. Under Balmer it absolutely had to be windows everywhere. Server, desktop, mobile, IOT, and all the rest. It wasn't working. That's why Balmer 'left' (i.e. was pretty much fired but in a nice face saving way).
Nadella killed that windows everything strategy and turned things around. Office is now liberated from Windows and making a lot of money for MS again. Cloud is growing more important as well, courtesy of mostly non windows software running on Linux in Azure. Meanwhile windows on the desktop seems pretty cool now that MS has added Linux integration, a non MS browser, and a few other things that make it a lot more attractive for developers and users and now that they are no longer forced to be in this windows everything walled garden. Even mobile is doing great for MS now that they can focus on treating IOS and Android as first class citizens.
- Apple was still on PowerPC processors until 2006 and weren’t really any threat to Microsoft.
- GSuite didn’t even launch until 2006 and even today, it’s really not that much competition to MS Office. It has less than a tenth of the revenue and only a few Fortune 500 companies (https://www.reuters.com/article/us-alphabet-gsuite-idUSKBN1F...)
- The iPhone didn’t come out until 2007 and really didn’t start gaining market share until 2010. Even today, Google’s “winning” with Android has only netted it less than $30 billion in profit since inception (https://www.theverge.com/2016/1/21/10810834/android-generate...)
- Microsoft adding WSL only matters to a few techies.
- Even the Nokia acquisition happened later during his tenure.
Windows Phone was a nice effort but years late and bogged down in a continued windows everywhere strategy and aggravated by lousy execution combined with institutional arrogance. They initially launched it as a windows CE shell with severe limitations. Then they alienated their OEMs and users with a highly disruptive and backwards incompatible move to windows NT. Finally, they topped it up with the Nokia acquisition. Nokia had committed early to windows phone which because of the prolonged uncertainty around this platform proved to be a major disaster. I used to work at Nokia, I've seen them fuck this up close. It wasn't pretty and Balmer was right at the center of it. MS ultimately acquired it and then proved it did not know how to fix it either. I still believe they could have pulled it off but not with the leadership (or lack thereof) that they had.
Balmer missed the boat utterly and completely, misread the market signals, and came up empty handed. The doomed Nokia acquisition was the proverbial nail in his coffin: the last in a decade long series of expensive failures.
The Google thing a few years ago was an existential threat to MS. Google was pushing users to chrome books running google office in a chrome browser on cheap hardware that ran Linux. The existential threat was two fold: no windows and no office. Nadella turned that around by making office 365 something that runs well everywhere, including on Chrome OS. The recent reverse takeover of Chromium by making that the centerpiece of Edge, sort of solidifies this. Up until then MS strategy was charging loads of money for windows on expensive laptops. Small form factor laptops became a thing during the mid 2000s and one key problem with those was that the windows license was substantial cost factor. It sort of made putting linux on hardware like that inevitable. It's the same problem they had on mobile. Expensive windows licenses for cheap phones did not make sense. That's why Android happened.
WSL indeed only matters to techies. But if techies pull their nose up for your platform, you have a long term problem because techies are the people that produce the things that make your platform long term viable. Five years ago, no self respecting engineer deploying on Linux would bother much with a windows machine. It was just too painful. Nothing worked and a typical README for developer tools on e.g. Github would be heavily mac/linux centric. Things like node.js, git, etc. barely even worked on it or only with a lot of workarounds and kludges. It was a big reason for Azure flopping in the market because MS sold windows in the cloud when the market was moving away from Windows. Nadella turned that around by embracing OSS, linux, and turning Balmer's mantra of "developers, developers, developers" into something that actually meant something to developers again. VS Studio Code and typescript now dominating the node.js and javascript ecosystem is quite a contrast with half a decade ago.
But if techies pull their nose up for your platform, you have a long term problem because techies are the people that produce the things that make your platform long term viable.
Techies don’t matter. The people that were making Windows viable were Microsoft themselves with Office, other major software vendors, and the dark matter developers writing bespoke software that never saw the light of day outside of internal use. There is a reason that VB6 still shows up in Stack Overflow rankings almost 20 years after MS abandoned it.
What were developers going to do? Ignore 90% of the market and wait for the year of Linux on the desktop?
IIS was used plenty for corporate Intranets and many low bandwidth sites. The dark matter developers who develop for the enterprise and go home at night and don’t post on HN or before that Slashdot are everywhere.
I worked at Nokia back in the day. Nokia knew this as well (I read the internal mailing lists, this was debated and speculated on heavily) and they did not act in time either.
2005 was in retrospect the key moment to act. Neither Nokia nor MS managed to do the right things. Apple and Google did. This was not exactly a secret except it did not translate into meaningful action.
Worse, Nokia actually shut down several projects that in retrospect could have saved their ass when it was handed to them by Apple in 2007 with the iphone launch. E.g. S90 (touch UI for S60) was killed in favor of S60 (no touch whatsoever until 2008) in 2005. Then in 2007 they went "oh fuck" and tried to bring it back on S60 ultimately resulting in a prematurely launched device that was everything the iphone wasn't (slow, ugly, unstable, etc.). Subsequent attempts to fix it yielded no results. Bringing in windows phone years later was an act of desperation.
Likewise Debian linux on tablets and phones was a thing in 2005 already. E.g. the N700 and N800 tablets shipped with it long before the ipad or iphone were a thing. Google ff-ing prototyped Android on these devices even before they shipped their nexus phone. Also, the chrome browser that made Android so successful? Guess what, Nokia had a webkit port for S60 in 2005 already. Google literally built Android/Chrome ecosystem on top of R&D done by Nokia (kernel drivers and optimizations, browsers, and other low level linux stuff).
I'm pretty sure MS has similar stories of death by management. Balmer was the key person in charge at MS throughout this.
2005 is also the same year that AJAX browser apps started displacing desktop UIs at a no doubt alarming rate for MS.
By the time MS was trying to push windows phone, their windows only development stack was a huge part of their problem. They needed developers to write apps for windows phone; especially developers already successful on IOS and Android. Only problem: all the IOS developers were on Macs (because Apple). And many of the Android developers were too (because they were pretty awesome). A windows laptop was a really hard sell to developers at the time because, well, they kind of sucked. Google was actively trying to compete with MS and still best friends with Apple so, the two of them created a developer friendly ecosystem outside of the MS bubble.
Because their developer ecosystem sucked, windows phone lacked apps, and therefore it flopped. Techies matter if you are a tech company trying to create and push new tech. If you are milking cobol, vb6, or some other ancient crap: fire all the techies. Basically, that's what IBM and Oracle did. Techies used to love IBM stuff. Not any more. Lets not talk about just how much they loathe Oracle at this point.
So techies not only matter but losing them cost Balmer his job and regaining their trust is a huge part of Nadella's current success.
Microsoft created AJAX - it was originally and ActiveX extension that only worked in IE.
Even today, GSuite is not that popular among big corps.
By the time MS was trying to push windows phone, their windows only development stack was a huge part of their problem. They needed developers to write apps for windows phone; especially developers already successful on IOS and Android. Only problem: all the IOS developers were on Macs (because Apple). And many of the Android developers were too (because they were pretty awesome). A windows laptop was a really hard sell to developers at the time because, well, they kind of sucked. Google was actively trying to compete with MS and still best friends with Apple so, the two of them created a developer friendly ecosystem outside of the MS bubble.
Developers go where the money is. It’s not the tiny Indy developers that are making most of the money on mobile. Windows Phone didn’t fail to get apps because developers didn’t want to program on Windows. They didn’t see any money in it. Heck even MS Office was available for iOS before Windows.
No, this is not fair.
Ballmer reconfigured sales channels, created an array of Enterprise offers including cloud - that's what enabled him to grow MS substantially during those years.
"Office is now liberated from Windows and making a lot of money for MS again. " - it always was.
The 'windows everything' strategy makes sense in the 2000's. iOS is pretty similar to MacOS - it worked for Apple. It even made sense early on in cloud.
Satya is the right choice for the right time, and it was right in hindsight to go cloud-linux and other things ...
But there's too much 'narrative' around this Satya thing.
When Nadella became CEO, they shipped it and it was a big deal. Nadellas chant was “Mobile+Cloud”. He came from Cloud & Enterprise. What he praised and asked everyone to focus on was very different from what Ballmer used to say.
Also - times change. In the 2000's people just thought MS was 'old'. Now they are realizing there's some life in the beast.
Satya deserves a lot of credit, nothing bad to say - but in hindsight it also seems as if MSFT was under priced.
But like the old saying goes:
“The stock market can stay irrational a lot longer than you can stay solvent.”
There was nothing on the radar between 2000-2007 that even hinted at Microsoft remaining dominant.
The one product that really sucks from this Microsoft is Azure. I commented in a thread before Azure is the absolute worst cloud provider amongst the big three after we thoroughly evaluated all three. AWS is the best and GCP is next.
But due to management pressure we had to go with Azure.
Unfortunately the best product does not always win in the enterprise. (That is why Steve Jobs never wanted to sell directly to enterprises).
The thing is that it seems to be the opposite when you're looking at it as a non-technical manager.
1. Terrible UI 2. Unclear Documentation 3. Frequent service outages. 4. Scalability issues, they were unable to scale a server in prod past 1000 connections where as in AWS/GCP they were able to 10x that with similar server specs. Something to do with how server defaults were set 5. Severe performance degradation and network exhaustion.
To be fair there were some positives but nothing exceptional.
But end of the day it kind of doesn't matter as Azure is going to be used in enterprises just because they have the legacy MSFT sales force that is tight with managements high up in the org. No wonder MSFT hit $1T.
Its such a shame for our industry though. Engineering teams at other cloud providers sweat and toil to get things right for their customers whereas MSFT uses other means to beat them. Hopefully they can at least fix things after they get in. But we know the chance of that happening is slim.
Meanwhile, Nadella is the man behind every single user-hostile and user-tracking feature you've seen introduced in Windows 10 (pretty much).
The question to ask might be, would Ballmer have done something else then focus on cloud? Like focus more energy on phones for instance? That could have led to a different financial result.
Point is that they still make the lion share of their revenue from milking PC OEMs and plumpy corporate clients.
The fact that they changed tools for doing so, does not change the scheme at large.
Their hosting service Azure just took over their AD and sharepoint clients with some extra being the genuine new clientele
I'd say that under Nadella, they grew even more recalcitrant, and dependent on their main revenue sources.
Ever got a visit from an MS salesperson recently? One that came to us to do AD was almost begging us to switch to AD on Azure.
I'd also say that MS became a little bit akin google in that their new product efforts fall into unending cycles of rebrandings, and half-hearted restarts. A lot of oldtimer MS devs I knew say that now they got a genuine fear of every new tech coming from MS almost customarily becoming an abandonware in 2-3 years term.
That’s how I have been feeling from around the time WPF came out...
And they reinforce that fear regularly.
That said, the board has done well picking leadership, Satya's driven a lot of amazing cultural change, and business too.
Care to explain? Do you work there? "doesn't work at all" meaning he's lazy?
source: worked at AWS 2008-2014.
But one thing people often forget is that AWS is cheaper than everyone else. This matters to most companies.
Also, Azure is a pretty miserable dev experience for many of their services. Many of AWS smaller services aren't great, but better from what I've used/general sentiment.
Can you provide a source for this please? I recently priced out AWS, Azure, and GCP for GPU workloads. I found that AWS and Azure were approximately on par, whereas GCP was significantly cheaper:
AWS p3.8xlarge: $12.24/hr [1]
Azure NC24 v3: $12.24/hr [2]
GCP n1-highmem-32: $8.311/hour [3]
[1] https://aws.amazon.com/ec2/instance-types/p3/
[2] https://azure.microsoft.com/en-us/pricing/details/virtual-ma...
GCP is, AFAIK, usually cheaper; but AWS is more established, and the “no one ever got fired for being buying...” player of public cloud (which has to chafe MSFT, since they pretty much displaced IBM in that role for on-prem.)
Alternatives for Exchange usually go in the direction of outsourcing (which ironically can mean: still Exchange, but not in the ball-and-chain on-prem way). Cloud options exist, but for people who for whatever reason require something to physically exist at a specific location (which is weird for email) commercial offerings are plenty (i.e. Zimbra, Zarafa, IBM's stuff, Red Hat's stuff, Oracle's Stuff).