Saving the Desktop
What do we do about the increasing surveillance built into Windows, Android, and MacOS? Hint: Linux won't save us
This week Microsoft announced its new line of Surface devices, which have a Neural Processing Unit (NPU). They’re building their Copilot AI into the Windows operating system. One of the new features is Recall, which records everything you do for later reference. Meanwhile, Google is testing a system they say will protect us from telephone scammers. It does this by listening in on all our phone calls. This has privacy experts… concerned.
We tend to take our phones and desktop computers for granted, but for years the tech giants have been boiling the frog (us) by increasingly monitoring everything we do. Back in 2019 we learned that Apple was employing external contractors to listen in on Siri conversations, and the system itself was easily triggered, potentially eavesdropping on Apple phone owners all the time. Microsoft has been criticized for incorporating extensive telemetry features in Windows 10 and Windows 11, which collect user data for diagnostic and marketing purposes, all of which set the stage for Recall. Since the introduction of Gatekeeper in macOS Catalina, Apple’s OS uses telemetry too. And it’s not just them—hell, even John Deere has used its software and DMCA laws to prevent farmers from repairing their own equipment.
Corporations using AI for nefarious purposes is such a common trope in sci-fi that I’m not even going to go there. It should be clear enough that, at the very least, Copilot- and Siri-like AIs could be hijacked by something akin to the attention-driven algorithms of social media, to ensure that the customer stays loyal to their platform. Your PC will soon be able to use every trick in the book, from flattery to guile to gaslighting, to keep your eyes only on it.
This is all very concerning, but there’s another dimension to it that you might not consider if you live in the United States. Chinese company Huawei has been banned from doing most business in the US, partly because of concerns about it taking over the American computer hardware business, but also because of worries that it might be building spyware into its systems. As this (now slightly old) Wired article points out, the temptation to put in such backdoors must be present in Microsoft and Apple boardrooms as much as it seems to be in the Chinese government.
If like me you’re not American, you are always aware that the desktop OSes you use (and most of your software) are built by US companies. Granted the political instability that appears to be gripping America right now, we in other countries find ourselves in the uncomfortable position of having to consider future scenarios where we treat American software companies the way we’ve been treating Huawei: with suspicion or even outright ban.
This idea may piss you off. Of course, Microsoft wouldn’t agree to turn Windows into a state surveillance system! Yet the movie Civil War takes the idea of a second American civil conflict seriously; after January 6, the world has to consider the possibility that US democracy could be hijacked. Because of this uncertainty, people who don’t live in the US should consider what we’ll do if something bad happens there. One of the most obvious questions is, what alternatives would we have to a computing and network infrastructure dominated by US software?
If you’re willing to consider the possibility of a civil war in America, then the idea that even a less-drastic swing to the right might result in Microsoft, Apple, and Google being required to operate as state instruments should seem equally if not more plausible.
Now Before You Rage-Click…
This is Unapocalyptic—I’m not in the business of making predictions, and in no way am I suggesting that Windows and MacOS, or ChatGPT, or American companies in general can’t be trusted. I don’t have an ideological stance on that either way; I would need actual evidence to point my finger and cry “j’accuse!”
What I am doing (as always) is building second-order or indirect impact scenarios—as Jerome Glenn pioneered with his Futures Wheel approach to foresight. If the first-order impact of civil unrest in the US is a government that plays fast and loose with its own constitution and international rules, then what are the second-order impacts in other parts of the world? If American software companies are put in the same position as Huawei, for whatever reason, how do the rest of us respond?
“There Is No Alternative.”
Margaret Thatcher’s favourite claim about neoliberalism seems to apply to the desktop, too. There are two choices: Windows, or MacOS. There is no alternative.
Yeah, there is Linux. It’s free, open (you can rewrite the code and compile your own version if you want), and it’s highly secure. It has plenty of desktop apps, many of which are better than their Windows or Mac alternatives. In a scenario where the world ceases to trust closed-source American operating systems, we’d all just switch to Linux, right?
Here’s the thing. Linux has been promoting itself as a friendly desktop environment since the beginning of the century. Nobody’s buying it, literally or figuratively. As a desktop operating system, Linux is a flop. Over the years endless articles have been written about why; they almost all miss the mark. Linux isn’t a bad desktop OS because there are too many distributions, or because it’s poorly marketed. Linux is a bad desktop OS because the desktop part is bolted onto a clone of a 50-year-old command-line-driven multi-user server OS that only a programmer could love. Linux, like UNIX before it, was designed and built to be used by software engineers. It’s completely intuitive for them. It fits the way they think so well that they can’t understand why the rest of us don’t love it too.
Lurking behind every friendly Linux desktop is a nightmare world of logic-bombs and software dependency traps that only specialists can navigate.
Every couple of years I dip my toe in the Linux pool. I install a distro, boot it up, and spend some time getting used to it. For a while, everything is great. It starts to look like I could use this thing as my daily drive. Maybe I could install it on my mother’s laptop and she could use it too. But inevitably, there comes a day when I want to do something simple—even something I’ve done a thousand times before—and I run into a brick wall. The system stonewalls me. Suddenly, I need the skills of a system administrator to resolve the issue—skills I don’t have. Now remember, I’ve been programming computers since 1980; I’ve written code in COBOL, APL, Fortran, BASIC, SNOBOL, LISP, and Perl. I’ve built my own PCs, configured domain names to activate websites, set up dual-boot systems, and hacked the Windows registry. I’ve worked in UNIX and VMS and mainframes, and was one of the first hundred or so users of the World Wide Web, back when it was a command-line program for accessing physics preprints at CERN. I’ve got loads of experience with computers—yet Linux always finds a way to defeat me.
It’s not that Linux isn’t ready for the desktop. It’s that it isn’t a desktop OS. No amount of lipstick on the pig is going to make it into one. (MacOS is also based on UNIX; the reason it works as a desktop where Linux doesn’t is because Apple tore out half of the UNIX functionality—like the multi-user stuff—and locked down every configuration option the user might use to accidentally break something. Doing this isn’t an option for an open-source OS; even if you gutted the multi-user complexities, who’s going to do the locking-down?)
Linux doesn’t cut it as an alternative to Windows and MacOS. We need an open-source desktop OS built from the ground up that can be used by non-technical people and international users.
A Promising Alternative
What we consider modern Windows was first released in 1993. Its chief architect was Dave Cutler, who had previously developed an outstandingly stable OS called VMS for DEC.
MacOS has a somewhat complicated history, but its roots lie in the NeXTSTEP OS of the early 90s. Then there’s Linux, which was first released in 1991. Windows is the only one of these OSes to be developed “sui generis”—by itself, with no precursor. The other two are derivatives of UNIX, which isn’t so much an OS as a dog’s breakfast of standards and utilities cobbled together starting in 1969. Most of the world’s computing is done on systems designed before there were graphical user interfaces, or a public Internet.
I will pass over OS/2 in silence. Wonderful as it was at the time, it faded away in the late 90’s—largely because Windows NT was superior in every respect.
Yet while the three operating systems we still use were being premiered, and while OS/2 was dying, another contender entered the field. This OS was designed from scratch to handle multimedia on modern, graphically-oriented personal computers. It was called BeOS. It was first released in 1995.
I could spend the rest of this post unrolling the whole saga of Be and its flagship OS; but you can find a good summary of that on Wikipedia. BeOS almost became Apple’s new operating system, but that deal fell through, and eventually Be went out of business.
BeOS lives on, however, in an open-source version called Haiku.
The Forever Beta
Many of Be’s developers believed BeOS was the Next Big Thing, and they had good reasons to. BeOS could do things in 1995 that Windows, Linux, and MacOS can’t do in 2024. (Try load-balancing the individual threads of your running programs, or switching processors on or off whenever you want.) BeOS was designed to run video, do 3d rendering, and juggle audio without a hiccup on 1990s hardware. Some of its capabilities are downright astounding—like the ability to drag UI elements of a program’s window and drop them into another program to create a hybrid interface. In the lower right corner of the Haiku screenshot above, the desktop switcher has been pulled out of its native app and docked as a ‘replicant’ inside the desktop itself.
The file system had metadata built into it, making it more database-like than other systems. The result: faster file searching and the ability to find and organize your files by their tags (video, audio, photos of grandma).
Haiku is essentially a clone of BeOS. It’s binary compatible with its predecessor. It’s taken a very long time for it to come together. It was only being worked on by a handful of people, and the non-profit created to develop it has only recently been able to hire its first full-time developer. Consequently, Haiku has been in beta now for twenty years. To say that this is a big problem is a huge understatement.
Signs of Life
Recently, though, new development tools have allowed developers to port Linux desktop apps to Haiku. This is huge—since Haiku conforms to the POSIX standard, programs written for UNIX or its derivatives are fairly easy to compile into native Haiku versions. You can see some of these in the screenshot above. Its default desktop is well established, clean, and intuitive:
A few weeks ago I decided to test it out.
Using the instructions on the Haiku site, I easily set up Haiku on a cheap, low-powered minicomputer I bought to keep up my Linux skills. Once I’d created a bootable USB stick, installing Haiku on this little box took about a minute. It boots in seconds and is blazingly fast.
A bootstrapping effect is happening: the more apps are ported over, the more people want to port more, and the easier it gets to do that. The more apps there are, the more people consider using Haiku as their ‘daily drive.’ There are 4000 at the time I’m writing. Now that it has several modern web browsers, I can even run Microsoft Office 365 apps in a browser window.
The OS itself has been rock solid, although it doesn’t recognize my sound card, and some apps have crashed. The experience takes me back to the early days of Linux distros, except that the core desktop functionality is spectacularly well-designed and consistent. Haiku is as easy to use as Windows ever was.
Having said that, a lot of the available software is more at the ‘technology demo’ stage than fully operational. Much more needs to be ported over, and the system will only truly come into its own when modern ported apps are rewritten to take advantage of Haiku’s unique features.
The biggest pain point is that it’s still in beta, after twenty years!
The Haiku team desperately needs to release Version 1.0. Depending on what happens in the November US election, I reckon they need to do that this year.
The world needs an alternative to Windows and MacOS, just in case.
LLM OS?
There is at least one project to build an open-source alternative to Copilot. Check out Fabric or Langchain and you’ll see that we’re rapidly entering a new era in personal computing; like Copilot and GPT-4o, these systems bring together swarms of Large Language Model agents, and using techniques like chain-of-thought they do real-world reasoning. But they do it on your computer, on your terms, privately, and transparently.
Rabbit’s ‘Large Action Model’ is supposed to give their AI the ability to browse the web, make airline bookings, and generally manage your life for you. Now that GPT-4o can talk to you (not in Scarlett Johannson’s voice), reason about pictorial information, and create images, spreadsheets, and much more, we have formally entered the age of AI sidekicks—computers we can casually converse with, and also ask to do things. Many desktop use cases go away in such a scenario because the AI can do the desktop or command line work, rather than us.
Microsoft’s strategy suggests that the desktop isn’t completely going away, at least not for a while. So we will want to layer an open-source LLM OS atop an open-source desktop. Right now, that means Linux; and right now, you need administrative skills, a knowledge of git and GitHub and of compiling programs locally to install such tools.
Most people have never heard of Github, and most people never will. What an average user needs is something that, like Copilot, just works. As I argued above, I don’t think Linux will ever be the platform where that can happen. Linux is solid and continuously updated, but its core architecture is not modern.
Haiku, on the other hand, is ideal for juggling complex multi-media tasks. It could easily be the foundation for a trustworthy open-source personal AI. And that would be truly empowering.
Wish List
For that to happen, Haiku has to mature. It needs to become stable enough for daily use (in my experience, it already is) but also be usable on a wide variety of modern systems, and that means more device drivers for, for instance, my sound card. The project needs more than one full-time developer; for this, they need money. This is where we can come in.
My own opinion, as a non-technical PC user, is that Haiku needs the following:
It has to get out of beta ASAP. We need Version 1.0.
People have to know it exists. Haiku needs publicity.
More tools and apps must be ported over, and those doing the porting should start forking application projects to take advantage of Haiku’s unique features. When the Haiku version of a platform like LibreOffice becomes better than its Linux equivalent, people will sit up and take notice.
The foundation for an LLM OS must be laid. This means building core logic into the OS for using NPUs, which is not something one person will be able to do. It would need a team effort.
The Haiku team has to act quickly. If they can’t release a viable desktop OS in the next year or two, Copilot and the other closed-source, commercial LLM OSes will be so far ahead that Haiku will never catch up. Merely being able to run Linux desktop apps just won’t cut it in that situation.
Act Now!
I urge you to head over to the Haiku site and, if you’re a developer, see how you can support it. The rest of us should try it, evangelize if we like it, and fund it so it can accelerate its timeline. For more information, see:
The main Haiku site, which links to everything. You can read up on it here, and download it to try it out.
A detailed review of Beta 4 (which is a couple of years old and rumoured to be about to be replaced by Beta 5).
Here’s YouTuber Action Retro asking whether you can use Haiku as your daily drive.
For the brave, the Nightly Images page of the Haiku site is where you can download the very latest (and potentially most dangerous) builds.
On the Donations page, you can financially support the project.
Give Haiku a look and let me know if you believe the world needs an open-source desktop operating system, in addition to the open-source server OS it has in Linux.
—K
I like Haiku but if it gets an LLM AI, I'll delete it and start over.
although i do highly support the arguments discussed here (without having read all of it i also have to admit btw!) personally, i don't care much about this privacy issue really even though i do strongly agree it is a very important issue ...
point is, there is also a good side to this 'tracking' performed on us online actually ... for example, it has also benefited me to find good articles like this one ... or find like-minded people and groups with similar or sometimes exact same interests ... but again, yes, it can be harmful in all kinds of ways and at times even scary in some respects ...
and as for different OSes out there, and why one might be more in use by many people than the less in use OSes, as a currently Linux (Ubuntu) OS user, i am really happy with it when comparing it with Window$, which hurt me for long enough, so as to make me do the migration eventually after nearly three decades of procrastination ... but still i do need Window$ or MacO$ etc once in a while because certain software (Photoshop namely) i do happen to need a lot once awhile are more readily and easily available on those platforms than they are on Linux or others ... (and let's not forget the fact that there are some other ancient yet not obsolete programs available on a supposedly 'dead' system, which i also need a lot at times ... neither of the well known and popular OSes can do that for me, unfortunately ...)