Qnx Self-Hosted Developer Desktop
Key topics
The QNX Self-Hosted Developer Desktop has sparked a lively discussion, with commenters weighing in on its potential to run on bare metal and its relevance in the automotive industry. While some speculate that QNX might eventually ditch QEMU and run on bare metal, others point out that it's already doing so in many cars, albeit alongside virtualized setups. The conversation takes a interesting turn when the cost of QNX for commercial use is brought up, with some suggesting that its expense is driving the shift towards Linux, which now supports real-time capabilities. As the debate unfolds, it becomes clear that the lines between QNX, Linux, and other RTOSes are blurring, making this development a timely and intriguing one to watch.
Snapshot generated from the HN discussion
Discussion Activity
Very active discussionFirst comment
23m
Peak period
58
0-6h
Avg / period
12.3
Based on 160 loaded comments
Key moments
- 01Story posted
Dec 26, 2025 at 8:16 PM EST
6d ago
Step 01 - 02First comment
Dec 26, 2025 at 8:39 PM EST
23m after posting
Step 02 - 03Peak activity
58 comments in 0-6h
Hottest window of the conversation
Step 03 - 04Latest activity
Dec 30, 2025 at 5:25 PM EST
2d ago
Step 04
Generating AI Summary...
Analyzing up to 500 comments to identify key contributors and discussion patterns
Want the full context?
Jump to the original sources
Read the primary article or dive into the live Hacker News thread when you're ready.
In theory I'd be tempted to try, in practice not, because of all the back and forth between changing owners in the past, and resulting policies regarding availability.
I'm also well served by some 'gaming distro', where nothing ever stutters or lags, on almost obsolete hardware, mostly clocked down to 800Mhz, with uptimes of up to 150 days. More isn't really useful anyways, because of updates.
But hey, Wayland! On QNX! with XFCE on top of that! Who would have thought?
[1] https://www.qnx.com/developers/docs/8.0/com.qnx.doc.neutrino...
[2] https://doc.rust-lang.org/rustc/platform-support/nto-qnx.htm...
The latter is actually a common setup, used by Mercedes-Benz and Hangsheng if I'm not mistaken.
(Running on Core i5 7500t and Core i7 7700t with integrated intel HD630 graphics on Lenovo M910q tiny with 32GB RAM. Mostly clocked down to 800Mhz. Chosen path: systemd-boot, Btrfs, ZRAM, Plasma/KDE)
You can already get a free license for QNX and grab a BSP (board support package) to create a bare metal image. You have been able to for quite a while. People who understand how a computer works, what a device driver is and how and when to use one, are not the target for this demo. It's targeted at the people who think the user interface is the software and the desktop GUI is the operating system.
And stopped.
https://hackaday.com/2017/05/03/your-next-desktop-qnx/
Granted, this is not the full Plasma shell, but you can run a lot of KDE software on it just fine.
They do list "A native Desktop image on Raspberry Pi" under What's Next, so hopefully soon:)
> In theory I'd be tempted to try, in practice not, because of all the back and forth between changing owners in the past, and resulting policies regarding availability.
Yeah, that gives me pause too. There was some noise earlier about open sourcing it; I do wish they'd actually do that.
The kernel processes are actual processes so each of the drivers is fully sandboxed, an error in one bit of code can not cause any other processes to be affected unless you explicitly declare that it should be so (shared memory, for instance) and of course you don't do that.
The reduction of scope alone is worth at least 30 IQ points.
Absolutely rock solid. I built some specialized network devices using QnX and those things ran for a decade+ after first installation. Not a single reboot.
[0] https://www.cvedetails.com/vulnerability-list/vendor_id-436/...
The reduction in scope is really gold, it makes it so much easier just to have a small defined interface per program. It is a bit like Erlang/OTP but with C as the core language, the IPC is so lightweight that it becomes the driver behind all library level isolation. So what in a macrokernel would be a massive monolith with all manner of stuff in the same execution ring turns into a miniscule kernel that just does IPC and scheduling and everything else is a user process, including all of the luxuries that you normally associate with user processes: dumps, debuggers, consoles.
Photon or not, I hated the period where they sort of moved to canned BSP deployment only, where in 6.5 I could just develop on a live system. This is nice.
In case you're not aware: CDE is still around, open source, and runs on modern unix-likes.
[1] https://github.com/fvwmorg/fvwm3
[2] https://github.com/NsCDE/NsCDE
> that doesn't require long time abandoned C code
https://github.com/NsCDE/NsCDE
Not so Common Desktop Environment (NsCDE) 2.3 Latest
on Jun 16, 2023
https://sourceforge.net/projects/cdesktopenv/files/
src 2025-11-25
CDE is still in active development. NsCDE is effectively abandoned.
Oh I'm aware :) also this beauty from SGI is now around again:
https://docs.maxxinteractive.com/
"This environment runs as a virtual machine, using QEMU on Ubuntu. To try the image, you'll need: Ubuntu 22.04 or 24.04." So it doesn't boot on bare metal?
Maybe they're trying to get away from needing Windows. The previous recommended development environment was cross-compilation from Windows.
The big news here is that they have a reasonable non-commercial license again.[1] The trouble is, QNX did that twice before, then took it away.[2] Big mistake. They lost their developer base. Support of open source tools on QNX stopped. As I once told a QNX sales rep, "Stop worrying about being pirated and worry about being ignored". They'll need to commit contractually to not yanking the non-commercial license to get much interest.
QNX should be licensed like Unreal Engine. If you ship enough products using it, it gets noticed and they contact you about payments, and if you're not shipping much product, Unreal doesn't care. This has created a big pool of Unreal developers, which, in turn, induces game studios to use Unreal. Unreal's threshold is US$1 million in sales.
Apparently they opened things up a bit last year, but nobody noticed.
Usefully, there is a QNX Board Support Package for the Raspberry PI, so you can target that. QNX would be good for IOT things on Raspberry PI machines, where you don't want the bloat and attack surface of a full Linux installation.
[1] https://qnx.software/en/developers/get-started/getting-start...
[2] https://www.theregister.com/2024/11/11/qnx_8_freeware/
Right. These days it's better to invest into Redox OS[1] as a potential substitute for it (if work on real time capability). And with real time patches merged into Linux mainline[2] QNX doesn't stand much chance today too.
[1] https://doc.redox-os.org/book/microkernels.html
[2] https://arstechnica.com/gadgets/2024/09/real-time-linux-is-o...
Correct me if I'm wrong but these and other Linux patches were always about soft real time and Linux never had hard real time capability because of its architecture.
That sounds quite a bit harder to enforce for an OS designed to run inside, often not internet-connected, devices.
I am a little confused about it both being a virtual image running on QEMU and requiring a very specific Ubuntu version as the base OS.
Shouldn't anything that can run QEMU work?
I used to do software development back on 386 with the OS on a floppy disk and really loved it. In fact, I ported my Objective-C compiler and runtime to QNX back then. Would love to play around with it on my Mac.
Thanks!
I’m sure it’s better now, it wasn’t so when QNX had come out.
I'd be honest and say that qnx demo disk had more usability overall than my disk, but one could easily have a usable text only linux bootable disk. Busybox and uclibc already existed back then.
https://www.cs.columbia.edu/~spotter/floppy.bin (won't be that useful today due to the ethernet drivers, but it was a 1.44mb floppy)
I wasn’t claiming Linux couldn’t achieve this, merely stating why QNX was mind blowing: because it was way ahead of what was available, not what was possible.
Congrats on your live floppy project, I guess.
my conceptual floppy was less to demonstrate it fitting on a floppy (that was just the carrot), the idea was to show how one could dynamically "rewrite" shared libs to only include the symbols one needed to run the apps and have them work.
argument being, static linking a single binary is smaller than dynamically linking it and including the shared library, but as one adds binaries, the duplicated code will eliminate that size advantage. But shared libraries (especially something like glibc) are large and apps dont always use vast sections of them), so what if one could iterate over all the dynamically linked apps one wanted to include and only include the sections of the shared libs that were needed.
So the project was demonstrating that. In practice, uclibc was smaller than the sliced up glibc (and at least for this project, worked just fine)
The page on https://devblog.qnx.com/about/ does not show what kind of company it is, who is behind it, and where they are located. Should I expect backdoors? Is it an elaborate front by north korea? Who will be able to remotely execute code on this operating system?
It's nearly 2026 and fake job applications by nation-state threat actors are common. If a new open source project with shiny marketing pops up it would really help if there is some proof that the org behind it consists of humans living in democratic countries.
Edit: The about page links to https://qnx.software/en which only shows a black screen for me.
[1] https://qnx.software/scripts/global/cookie-consent.js
Stop sowing FUD https://en.wikipedia.org/wiki/Fear,_uncertainty,_and_doubt
Yes, it is. It's specifically called concern trolling: https://en.wiktionary.org/wiki/concern_troll
In this case it's because the mass-market operating systems with which QNX could compete already do the things you're “concerned” about. QNX could only be an improvement in that regard.
As someone who has not worked in automotive QNX is a totally unknown brand and based on the linked website I had trouble finding out what it actually is. Also the wikipedia source just stated RIM and might not have been updated at all. Also RIM/Blackberry is not a brand that is positively recognized.
With something like Ubuntu it was an easy, verifiable story who is behind it and what they are doing. That's what the linkd QNX page was missing, and I pointed it out because I actually tried to do the due diligence and see what company is behind it and where it is located.
That said, like IBM and Microsoft, they've also been on and off over the years about whether tinkerers, desktop, and other uses are welcome. So they probably could benefit from showing that this time they're opening the ecosystem for the long haul.
https://www.theglobeandmail.com/investing/markets/stocks/BB-...
I know it's a microkernel which is inherently cool to me, but I don't know what else it buys you.
Can anyone here give me a high-level overview of why QNX is cool?
It's also really compact. This used to be a great selling point for underpowered car infotainment systems. Some cars had around 1Mb of RAM for their infotainment, yet they were able to run fairly complex media systems.
QNX is also used for non-UI components, just as a good realtime OS.
They were also early adopters of Eclipse, which was the "default IDE" before the advent of VS Code.
Sadly not anymore, Qnet was removed in 8.0
In general QnX was commercially mismanaged and technically excellent. I'm imagining a world where they clued in early on that an open source real time OS would have run circles around the rest of the offerings and they'd have cleaned up on commercial licensing. Since the 80's they've steadily lost mind and marketshare though I suspect they'll always be around in some form.
I've also noticed that all of the message passing system calls still accept the node ID. Are there plans to open up this interface to allow for custom implementations of Qnet, maybe? I'd be very interested in exploring that.
[1] https://www.qnx.com/developers/docs/6.5.0SP1.update/com.qnx....
[1] https://www.wincustomize.com/explore/litestep/6/
[2] https://www.wincustomize.com/explore/litestep/292/
At the end, I had really beautiful (to my eyes, back then) and very functional desktop, but something went wrong when I made backup before installing SuSE Linux 7.0, so months of vigorous customizing were lost. :-(
But it was fun while it lasted. There were a number of alternative desktop shells in the Windows 95/98 era.
I have ever seen.
https://github.com/swiftlang/swift-testing/issues/868
We run into a lot of OEMs who switch to Linux because of AGL and come crawling back to QNX many expensive months later to start over with a viable solution so they can deliver.
QNX hypervisor architecture allows companies to adopt AGL gradually bringing in the productivity benefits of stuff like flutter and Linux. QT is hell, their entire moat has been the automotive industry/highly embedded devices and seeing Flutter chip away at even a small portion of it gives a lot of hope for a better competitive landscape for more productive and performant embedded guis.
Some comments mentioning QNX can run Swift code makes me think of it could also run iPhone apps.
While Blackberry exited the phone market, I’m surprised to know QNX is still the most popular os for cars. With 275 million devices running it atm.
Not at all. That is like saying because it can run C, it can run windows apps. To run iPhone apps you would need all the libraries and runtimes ported, including the whole GUI. Just not happening.
It's great that Apple are pushing Swift out there a bit, but honestly if they want the World to catch fire with it, they need to give away the Crown Jewels and get SwiftUI out there as well.
Meanwhile, it's great that QNX is supporting modern languages. I can imagine having a bit of fun with this developer desktop and seeing how modern tooling plays nicely with it.
On at least one occasion, the license was changed overnight leaving a large enthusiast community in the lurch.
Given the history, there's every reason to suspect that there'll be yet more rug-pulls in the future.
And now, for the 3rd time, the are offering “free” license.
I think it's a real shame Blackberry didn't manage to etch a third (or fourth - I also loved Palm's WebOS) niche for their QNX-based phones. Blackbberry 10 was an amazing mobile OS.
100% this. I had a Passport and it was one of the single lovelist phones I've ever had.
Compared to my Nokia 7710, the last device with the original Psion UI... that was an elegant touchscreen, plus physical buttons, and a replaceable battery, but that was about it.
Compared to my Nokia E90 Communicator...
The keyboard was even better; it charged off a standard MicroUSB port, and had a standard headphone jack; it had way more apps, because it ran Android ones pretty well.
Compared to any Android phone... Vastly unrecognisably better messaging app, with one inbox for all messages and notifications. Square screen so no fighting portrait vs. landscape. Physical keyboard for much more accurate typing -- and scrolling. Google-free.
I would still love one, but I don't think I could move it to my own Blackberry account at this point in time.
it run qt and does everything else so it is often an easy choice.
Initially the actual implementation didn't match the conceptual framework, but by version 1.2 they had really cleaned things up.
QNX will shift focus in a year or two.
5 more comments available on Hacker News