I’ve been feeling gushy about my setup lately, I think I’ve finally found my home on Linux. For decades I’ve distrohopped each year and never was really happy with it all, but Fedora Atomic has changed that.
Some things I can do with Fedora Atomic that I cannot do with other Linux distros:
-
I can rebase to Bazzite for gaming performance when I feel like having a long gaming session.
-
I can rebase to Secureblue when I think I will not be gaming and would prefer a more secure linux setup.
-
I can update my system and not have to worry about special instructions, its extremely stable. Many times in the past, running a small ma-and-pa distro with most things pre-configed for performance would end with it breaking after a couple of major updates. This isn’t true for configs like Bazzite and Secureblue, they are remarkably stable across many major updates due to how rpm-ostree functions.
-
Distrobox and Flatpak are more than enough at this stage for most programs and they help you avoid making too many alterations to the base image, greatly speeding up the swaps between major images.
The kicker? Your user configs and home files are never changed when you ‘image hop’. It always feels like you just installed a fresh distro whenever you upgrade, and the performance benefits are noticeable. You don’t have to tinker and do the same changes over and over, its all handled for you by rpm-ostree.
10/10 this is the future of Linux. I hope for a future where I can rebase entire Linux distros while maintaining my configs with one simple command, but for now, Fedora Atomic is fantastic.
The downsides:
-
There is one major downside, and its that all of your system files are read-only. Personally, I’ve found a dozen ways to get around this, it requires thinking inside the Distrobox. It is a notable issue for many people, though. This means you cannot make specific tweaks without making a whole new image for yourself. Though in practice, I have found the ecosystem has grown a lot. Other people have already made the best tweaks available for you with only a few simple commands.
-
Rpm-ostree also is slow to update because its essentially building a whole git tree to make sure your updates never break and are as stable as possible. You also have to reboot each time you alter it, which can be annoying, but if you stick to flatpaks and distroboxes, this issue is mitigated significantly.
Totally get why you’d ask if you’re the only one on openSUSE MicroOS, especially with all the buzz around Fedora Atomic. Let’s explore what the latter has going for it that have helped their adoption race ahead:
Head Start & Delivering Desktop Variety: Fedora Atomic desktops’ efforts simply got rolling earlier. Importantly, they also managed to deliver a solid KDE Plasma option (Kinoite) alongside their GNOME flagship (Silverblue) in a reasonable timeframe. For instance, the ideas for Fedora Atomic started around April 2014, and Kinoite hit beta by November 2021. Now, consider openSUSE: their work in this immutable space (with some roots in Project Kubic around May 2017) is still, as of May 2025, working towards a beta release for Kalpa (the KDE version). This extended wait for a polished KDE experience – a desktop environment hugely popular within the openSUSE community and beyond – undoubtedly has implications for overall adoption and even the perceived momentum of MicroOS as a desktop project. When a major DE option, especially one with KDE’s broad appeal, is lagging, it can slow things down.
The uBlue Phenomenon: Can’t stress this enough – community projects like uBlue (and its offshoots) have been an enormous catalyst. They make Fedora Atomic super accessible with pre-configured images including NVIDIA drivers, codecs, and common tools. They’ve likely single-handedly tripled or quadrupled the user base for Fedora Atomic by just making them work out of the box for more people.
Fedora’s Broader Reach: Generally, Fedora just has a larger overall user base than openSUSE. While both are fantastic, top-tier distros, Fedora’s wider existing audience naturally gives its specialized spins, like Fedora Atomic, a larger initial pool of potential users to draw from.
So, what’s the deal with openSUSE MicroOS (or Aeon/Kalpa for desktop)?
Right now, its approach to being “managed” or declarative feels a bit milder. When people dive into immutable/declarative systems, they often see a spectrum:
People often move between these. If NixOS is too much, Fedora Atomic becomes a common landing spot. If they love what Fedora Atomic offers but crave even more control, they might look to NixOS. For someone already in this mindset, openSUSE MicroOS’s unique draw isn’t as sharply defined yet. And let’s be real, for many long-time openSUSE aficionados, the fact that YaST – arguably a killer feature and a huge USP for the traditional distro – isn’t really part of the Aeon/Kalpa experience (or MicroOS generally in the same way) definitely stings a bit. It feels like a missed opportunity when such a cornerstone tool doesn’t quite make the jump to the new paradigm.
Where I think openSUSE MicroOS Desktop is compelling is as a super logical next step for openSUSE Tumbleweed users. It’s less about being an entirely new, radical thing like NixOS (or even Fedora Atomic in some ways). It’s more like it takes the best bits from atomicity and transactional updates (think easy rollbacks with
transactional-updateand a read-only root) and blends them into the fantastic Tumbleweed foundation.So, it’s an evolution of a trusted model, beefing it up, rather than a completely different animal. This is probably the openSUSE team’s vision. Time will tell how it fully distinguishes itself, but it’s a smart way to get modern robustness without throwing out all the familiar openSUSE goodness.
Hope that lands better! It’s definitely a space changing fast.
Why are you using ChapGPT to make comments on lemmy? What’s the point?
In this case, I woke up in the middle of the night. I couldn’t sleep. Saw this and wanted to answer somehow. Went to use speech to text for a first draft, you may find it below. As I’ve got pains related to RSI, however, I couldn’t be bothered to make it all slick and crisp myself on a phone. My laptop + split-keyboard setup was tucked in my bag. So, at that moment, I asked an LLM -unsure at the moment if it was trained by my own data to replicate my style of writing- to perform the ‘act’. After some back and forth, we got to the final result. Content-wise, I’d say, it’s all just me. The LLM only did wording/phrasing/formatting etc.
(The original draft from speech to text:)
Due to the order of how events have happened, i.e. the fact that Fedora Atomic matured earlier, simply by virtue of being earlier into development, and also because the idea to make a desktop out of it wasn’t just an idea that was tagged along later, but an important thing a lot earlier into its development These are definitely key reasons for why the adoption of Fedora Atomic has been a lot better than OpenSUSE micro OSes And I haven’t even mentioned the fact that a fan project like Universal Blue has had for the adoption of the ladder Heck, it’s easy to sum up in retrospect, simply because the data is there, that Universal Blue has single-handedly, maybe tripled or quadrupled the userbase of Fedora Atomic, hence all of the above has helped Fedora Atomic’s adoption a great lot Of course, Fedora is, for some reason, more popular than OpenSUSE, while they are mostly just different continental ideas, or distros, rather of the same idea, or close enough Regardless, as to your question regarding OpenSUSE micro OS, I think that with the way they’ve set it up It is relatively mild, at least at this point in time, to managed-ness
and abiding to the rules of congruent system management which means that if someone likes what Federal Atomic does in this regard as it is, at least in this point in time, by far the most popular of the Atomic branch of together with NixOS they often switch between these i.e. if NixOS is just too hardcore or its language is just a little bit obtuse for what they want out of the system then its easy for them to just simply adopt Federal Atomic instead or if they like Federal Atomic, what it is, but want to increase the level of managedness and going full declarative, then they can go for NixOS instead but having started from either of these, the unique selling point for OpenSUSE microOS Desktop is simply not there yet, or at least not as pronounced as it should be as for what I think, OpenSUSE microOS Desktop seems like a very logical step up from OpenSUSE Tomb Raid, which is probably how they envision the project at least if we would ask Richard Brown of course time will tell if the one will go over into the other or vice versa regardless, it is more interesting, in my opinion, as an evolution of the traditional model that adopts the most minimal of what atomicity and transaction updates has taught us
Rather than being a new paradigm in its entirety that tries to do or be as radically different as either Fedora Atomic or NixOS tries to be.
Do you always have ideas in the middle of the night and want to post them only to have an RSI flare up and no laptop nearby and decide to use ChatGPT to write your posts?
It’s not just this response. All of your posts read the same way.
Like using AI as a writing assistant is fine and all. But the posts you copy paste over are mostly LLM structured arguments.
Strictly speaking, for my posts[1] (i.e. my comments aren’t included into the conversation yet), I do heavily employ LLM as a writing assistant. But the process those undergo is very different from the comment you see above; it takes a lot of time, effort and many revisions until I land on something I like.
As for my comments, it depends: if it’s longer, I employ it to help with shortening while retaining the content I meant to convey. Or, to help with wording/phrasing specific troublesome passages that either don’t flow well or if I’m unsure if idioms (and whatnot) have been used correctly.
While I don’t like to bring it up, some people -naturally- have the tendency to write up texts that are (somehow) reminiscent to what we’d expect from an AI. FWIW, I have many times been accused of this while the text was all just me…
Finally, to directly address the comment found above: No; I don’t think I can recall any other comment that was as carelessly composed as that one. And to directly answer your following question:
Nope. I can’t recall the last time -prior to the one above- in which I did something similar.
Which there are only three of at the time of writing. ↩︎