Ubuntu 26.04 Lts – the Roadmap
Posted5d agoActive3d ago
discourse.ubuntu.comTech Discussionstory
informativepositive
Debate
20/100
FreebsdLts ReleaseLinux
Key topics
Freebsd
Lts Release
Linux
Discussion Activity
Active discussionFirst comment
1h
Peak period
16
6-12h
Avg / period
6
Key moments
- 01Story posted
Dec 28, 2025 at 5:13 AM EST
5d ago
Step 01 - 02First comment
Dec 28, 2025 at 6:38 AM EST
1h after posting
Step 02 - 03Peak activity
16 comments in 6-12h
Hottest window of the conversation
Step 03 - 04Latest activity
Dec 30, 2025 at 12:09 PM EST
3d ago
Step 04
Generating AI Summary...
Analyzing up to 500 comments to identify key contributors and discussion patterns
ID: 46409976Type: storyLast synced: 12/28/2025, 2:55:53 PM
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.
I want my OS updates to be boring. Granted I'm using Kubuntu (Ubuntu with KDE) so the Gnome stuff has nothing to do with my use, but the fact that there is nothing there that I have to fix or anticipate or work around or develop a new workflow for is terrific. That's what I love about the Ubuntu family - the last time I had a major upheaval with my desktop system was the year after KDE 4.0 was released... I think over a decade and a half ago. I really have not had to think about my desktop since.
https://news.ycombinator.com/item?id=46412188
[1] https://ubuntu.com/about/release-cycle
[2] https://wiki.ubuntu.com/NobleNumbat/ReleaseNotes/Kubuntu
- Easy access to (newer versions of) some particular package in distro repos
- You prefer the Kubuntu default experience more and don't want to customize
- Company policy
I've officially missed a whole cycle!
jkjk, thanks for the hard work, I'll wait as long as it takes.
The alpha was in early September:
https://www.theregister.com/2024/09/12/pop_os_2404_cosmic_de...
The beta was the end of September:
https://www.theregister.com/2025/09/30/pop_os_2404_beta_rele...
The release date was announced as mid-December in November at the Ubuntu Summit:
https://www.theregister.com/2025/11/03/cosmic_1_before_xmas/
The full final release shipped before Yule:
https://www.theregister.com/2025/12/22/popos_2404_cosmic_epo...
If you care, were you not paying attention since the summer?
I get what they're going for - a way to ship self-contained (usually end-user-facing) applications with any dependencies they need without any risk of breaking other applications in the system. Unfortunately, it just results in breaking those applications specifically instead, in weird and stupid ways that are difficult to debug.
I think if snaps did the Flatpak thing - extract to a local directory instead of living on squashfs forever, or even storing them as an uncompressed disk image instead of squashfs - it might be more reasonable, but at that point you may as well just use Flatpaks like everyone else wants.
[0] - Add the following to `/etc/apt/preferences.d/no-ubuntu-firefox`:
Then install the apt repository as described here: https://support.mozilla.org/en-US/kb/install-firefox-linux#w...This will make any `firefox` package from any repository with the `Ubuntu` label (i.e. an official Ubuntu repository) have a -1 priority, or 'never install ever'.
I'm not a Debian expert, but isn't apt used for this ?
Similarly to rust coreutils, fake sudo and the likes that they push.
But many people will never pay for Linux and it's even causing people to move away (eg to Mint which removes snap)
Perhaps it makes sense in the enterprise market though. They're always trying to push launchpad to us at work and I'm sure this will integrate with snap. But launchpad doesn't work for us because it only works with Ubuntu. So it's just a non starter for us, we have more distros to support. Sure Ubuntu is the biggest in our environment but we want a single pane of glass for everything. More similarities between distros would make that a lot easier.
That, and in case of snap, is to create artifical market share for their proprietary and paid solutions by preinstalling it on the free version.
That said, I find myself increasingly at odds with the direction they're taking. The whole Snap vs. Flatpak debacle is exhausting, and personally, I'm not a fan of either. I'd take a standard apt repo over containerized desktop apps any day. Seeing core applications migrate to Snaps and the recent decision to move coreutils to alternate implementations feels like a bridge too far for my taste.
There's also the creeping Proprietary integrations to consider. To be honest, this is more of a philosophical stance than a practical one. Ubuntu is still a fantastic "get work done" distro, and I still use it on my office laptop because it just works and it's the only destro that got my employer's stamp of approval.
But for my personal setup? I've moved on. It's Arch for the desktop and Debian for servers. Nothing else really hits that sweet spot of control and simplicity for me anymore.
It's a classical embrace and extinguish strategy.
I personally welcome this change as the current GUI is indeed a bit confusing.
> laying the groundwork
So with constant focus, how many more years before the feeling is reached on top of that groundwork . The map is rather fuzzy
I used this for a long time and still do sometimes. However, Arch works well enough now that I don't need to bother with Windows anymore. It is much more efficient for working with containers as there is no VM involved.
As an especial what the fuck are you doing, for the LTS 24.04 release that nvidia tested against, canonical decided to upgrade their kernel, without bumping their minor revision number, to one that cuda doesn't run on. Downgrading that to the kernel 24.04 originally shipped with broke zfs, which Ubuntu made a huge fuss about shipping out of the box.
Damned thing is running now (without zfs, and gnome won't start), and I think I've killed the automated updates system, but it definitely doesn't have robust wont-fall-over vibes.
So for Canonical, if you see this, don't change the kernel you've released with if you aren't also changing the version number.