Merged a settings page in 1.10.0!

New AI policy update: LLM-generated content is no longer allowed for code, comments, or descriptions https://github.com/pop-os/cosmic-epoch/pull/3911

The only code contributions that we have approved that used AI assistance so far have been minor fixes and improvements where the risk of regression and copyright violation is low. The requirement to certify contributions under the Developer Certificate of Origin also makes the person submitting the changes liable if they are submitting copyrighted code.

Any of the more complex changes and feature developments that people have attempted have been denied and will continue be denied. If you look through all of the changes that we do merge, the vast majority of large feature developments and fixes are done by hand by trusted contributors (ie: Hojjat) and official team members who are consistently and frequently contributing large changes to COSMIC that did not involve any LLM use.

It is more the exception when a contribution that used AI assistance was used that is able to be approved and we will likely make some policy adjustments due to the amount of burden that they’re currently causing in the review process. Truthfully the only situations where AI has been useful over the last year is when the person used it for debugging purposes rather than code generation.

Accepting some contributions that used LLM assistance and being pro LLM are two entirely different things. As per the AI policy[0], use of LLMs is heavily discouraged. Contributions that fail to check every item on that list will not be reviewed. Those that correctly document their use of LLMs in each commit description will be heavily scrutinized with an unfavorable level of trust as the baseline. This filters out most LLM contributions so that only the highest quality contributions are reviewed. Lower quality contributions will likely be rewritten by hand in a completely different way if it solves a problem.

[0] https://github.com/pop-os/pop/blob/master/.github/PULL_REQUEST_TEMPLATE.md

It is normal for cosmic applications to write logs to the terminal.

You have to type the key sym name into the input field. It is not a raw shortcut editor yet.

You are already on the LTS version if you’ve installed the COSMIC Alpha 6 ISO, or did a release upgrade.

You can choose a session in cosmic-greeter through a dropdown.

You don’t have to watch. You can listen to it in the background.

It’s really easy to configure a self-hosted forgejo instance. Even if you rm your local work, you can clone it from your server. Be that hosted on the same system over localhost, or on another system in your network.

3D acceleration is required with cosmic-comp right now. I’m not sure if software rendering will be ready or not for the first release, but it is on the issue board.

There’s already been explanations in every thread on COSMIC for the last 2 years. Along with a dozen interviews and conference talks. Why are you demanding answers here?

See the Ubuntu Summit 2024 talk: https://www.youtube.com/watch?v=bwrBKccfYws

It’s not resources, in fact, this Alpha performs pretty poorly on its own vs Gnome

I haven’t seen any benchmark where GNOME was more performant than COSMIC. Despite alpha status, it is already much more responsive than GNOME.

GNOME uses a single thread to render all displays in a multi-display configuration. This is often so slow that they need to rely on double or even triple buffering when the frame rate lags behind the display’s refresh rate. Meanwhile in COSMIC, thanks to the thread safety features of Rust, it was easy to implement thread-per-display multi-threaded rendering. This means that each display is rendered and composited independently on their own respective threads.

GNOME’s compositor also has an entire JavaScript runtime bundled inside of it, which it uses for drawing interfaces and handling application logic for those interfaces. All within the same process as the compositor, slowing down its event loop. COSMIC instead keeps the compositor process very lean, with all desktop interfaces running in their own isolated processes outside of the compositor via wayland’s layer-shell protocol.

If you can’t see the difference, it’s because you’re not even looking.

It can’t be fixed without forking and rewriting a lot of gnome-shell’s internal logic.

Also, COSMIC is not a rewrite of GNOME. Not even close. It is a completely different architecture, toolkit, language, and design system.

deleted by creator

Doesn’t matter. New compositor: cosmic-comp. Does not share any code with gnome-shell or mutter.

OpenGL is required, even if by software rendering.

No modifications are being done to Qt/KDE theming, and we won’t be able to get KDE apps to adapt to the COSMIC theme until the KDE ecosystem has finished migrations to the Union theme engine.

I don’t know how you can keep telling me that I never contacted Canonical even though I did. Nor did anyone ever publicly mock Canonical. You are putting words in our mouths. So much contradictory and hyperbolic nonsense here. Let me guess: you read a certain hyperbolic hit piece from a Chris Davis—one of the most prominent libadwaita and stopthemingmyapp developers—whom had a personal axe to grind with us because of many heated online arguments with him over the petition, theming, and libadwaita. He created a hit piece to influence public perception of the company and intentionally used the GNOME blog to reach the widest audience for his vendetta. Even though if you dig through the details his statements are weak, if not outright false. To make matters worse, GNOME never addressed that personal blog post hosted on their website, even though we had been sponsoring and sometimes hosting GNOME events for 5 years prior, and donated a total of $100K. Leading many to conclude that this was the voice of GNOME, even if internally it was not. This is what happens if you only read the story from one side without putting equal weight on the other.