Not a clue myself, but when I have to choose, I go with Mistral. They used to have a very good policy last time I’ve read it. They’re also targeting the European market a lot (and are located in Europe), which seems much safer than US or China in terms of data handling to me.

They also shared open-weight state-of-the-art models and ideas multiple times. The 7B Mistral was a thing for a while, and they were the first to ever publish an open-weight MoE model (Mixtral) I believe.

Unfortunately I don’t think I know the answers to your questions. I do what you say: open one or multiple tabs in my terminal emulator if I need to read multiple parts at the same time. I press / and then type in some case-sensitive text if I want to search for occurrences of this text. Or I press ? and some text to search backwards. And the rest I “lamely” do with pressing up-down arrows…

Learning by example is exactly the second layer I’m referring to in my text above. Man pages are very… formal. If I don’t yet understand what they even talk about, I need examples before their concepts become “native” to me.

I can’t “teach” well really - but I’m explaining how this works for me. I used to hate man pages, now I love them. Even a very difficult one like man bash. There is a point in the learning curve where it gets a lot better…

To OP or others, please re-post if you’ll do it through epetitionen.bundestag.de!

Yes, that’s the best way to look at it! :)

Come to think of it, I’ve been using Linux since about 2006. So I guess I’ll celebrate 20 years of Linux on my Desktop :)

Great! Would be happy to learn more about that in the future as well

If thinking about zed, Gram is a mandatory mention. https://gram.liten.app/

I just want to point out that you wrote your comment after I’ve already edited mine, where I’ve clarified that I misread the text in the post.

I’ve simply misread. Please don’t add too many constructs about AI on top. It was easy to misread IMO too, others have done as well.

Thank you for the clarification! That makes more sense honestly. Indeed I have read it differently at first.

This seemingly was a mistake with Linux’s second-in-command Greg Kroah-Hartman responding to each report asking about their submission to the public mailing list.

I hope that was not AI automation, with him being a lover and a loud public advocate of it.

EDIT: we’ve misread the text. It wasn’t a mistake by Greg, but merely that he was the one to reply. Nobody directed Trend Micro to the public mailing list, they did it by themselves (as far as we know).

• I think AI disclosure has to be a hard rule. Not done properly? Remove.
• I think there has to be a threshold of when to allow, otherwise we’ll get (and are already) overwhelmed. For example, be X months old is an easy check that would filter out low-effort posting.

Besides these basic rules that I think are absolutely mandatory, I’m not sure how far to go. I’d want to say that the project needs to explain Free Software affinity or explain why it’s posted on lemmy or something like that… But it’s so easy to fake it with, the irony of it, AI texts.

In conclusion,

• Decision-wise, I wouldn’t mind if the resolution is the full removal on AI-driven projects.
• Decision-wise, I would absolutely request a hard-ban if the lemmy text is AI-written.
• Decision-wise, I could accept if the rule is to accept some small sliver of AI posts that would come our way otherwise. Maybe some projects where AI assists in writing code, but does not replace the programmer. Maybe that is OK, if the project otherwise is “relevant to the open source ideology” (Rule 1 of this community).

EDIT: Also, something to think about is… community, not just code. For community health, it might be better to remove AI-driven projects entirely. Because we aren’t just about code. We’re also about continuing to exist long-term.

EDIT 2: I think codeberg actually did it better than my improvised thinking. I’d go with codeberg’s rules: https://blog.codeberg.org/protecting-our-floss-commons-from-llms.html

👍 Makes sense. Also, thanks for the additional info. (“the Python devs are saying to update pip and uv”)

Oh, your comment made me understand the source of the confusion, I think. I was speaking about the distro itself, hinted by the author above me using the word “distros” as well. But if we’re speaking about pip/npm/etc, then reverting those does make sense, even if a bit paranoid (but being paranoid is no problem). I’ve edited my message to help resolve this misunderstanding.

What good would rolling back updates do? Also, most if not all distros use package signing. It’s by far safer to update forward, not backwards.

The advise to “rollback” seems actively harmful, to be honest.

UPDATE: I think I understood the source of the confusion. If we’re speaking about rolling back the distro itself, then that would be actively harmful. On the contrary, *do* upgrade ASAP and don’t roll back. However, if we’re speaking about stuff that is derived from the system, such as pip/npm/etc registries, then latest artifacts from those could be compromised, so you could roll back those to instead get fresh-er upgrades from an upgraded distro where the signatures are already removed. I think that explains our contradictory-at-first-glance views. If the above is true, we’re actually in agreement on what to do.

Could you please explain a bit, do you mean that otherwise your bluetooth device would not auto-connect to both devices?

I mean, you probably can’t use them simultaneously anyway?

No, I mean the drivers for the graphics card. Xorg is a layer higher than that…

At this stage, and for the pioneers in changes like that, just the publicity around it is already a huge deal. I’d say it’s worth it even as it is, already. If the trend will go stronger into more and more adoption of these free software solutions, then we accordingly should push and expect more compensation to the underlying technologies and communities.

Would be proper to attach some links for such a claim, I think. (I’m not saying you’re wrong, I’m just saying a link would be proper.)

Indeed. I don’t know gentoo, but from the logs it seems that the failure somes from not being able to load the drivers. Install them first?

Good day Arthur, I think it makes sense to respect lwn’s publishing policy and rather publish the non-subscriber link. It will be available to non-subscribers in 6 days: https://lwn.net/Articles/1099594/

Writing this in case you haven’t noticed this detail.

It’s a dangerous line to consider. What if some subsystem or person does not have AI? Will it now become mandatory? Which exact model will they use, and is that model even free (as in, training data available)?