ProdigalFrog

6

The reasons it's a problem (spoiler, click to expand)

The Licensing Problem:

Unknowingly using potentially copyrighted code from an LLM in a FLOSS project opens it up to being sued for infringement, which most FLOSS devs can’t afford to fight, especially in the US’s current extremely pro-corporate courts.

It’s putting a target on your back for down the road when it becomes profitable for patent trolls to use AI to try to scan for copyrighted code on public code bases. Big tech companies could do the same to squash an open source competitor.

Both Haiku OS and BSD are not allowing AI code contributions for this very reason.

The Ethical problem:

Simply using an AI that's run on a corporate data centers encourages the construction of yet more data centers, with all of the environmental/climate negatives they bring, as well as local harms they induce on the people living near them, such as increased electricity rates.

Using corporate AI directly helps the financial situations of the big tech companies that host them (by boosting usage/user numbers, they are able to attract more investment capital), most of which are owned by right-wing CEO's who are more than willing to collaborate with and fund fascist governments to ensure that they are not regulated in search of both maximum profits. Some of these companies, such as Nvidia, Palantir and Oracle, genuinely appear to be seeking to use these tools for what would previously be considered crackpot conspiracy theory levels of public control and surveillance.

But if someone is absolutely determined to use them regardless, I would hope they either use a locally run LLM, or use a distributed model like AI Horde, as at the very least that does not encourage the construction of corporate data centers nor boost their financial prospects.

ProdigalFrog @slrpnk.net 1mo ago

I agree, that would be good to know.

ProdigalFrog @slrpnk.net 1mo ago

The problem is both using LLM code in an open-source project due to the threat that poses to the project if they unknowingly introduce copyright infringement, as as well as the very real ethical concerns inherent to using corporate owned and hosted LLM models.

The problems in detail (spoiler, click to expand)

The Licensing Problem:

Unknowingly using potentially copyrighted code from an LLM in a FLOSS project opens it up to being sued for infringement, which most FLOSS devs can’t afford to fight, especially in the US’s current extremely pro-corporate courts.

It’s putting a target on your back for down the road when it becomes profitable for patent trolls to use AI to try to scan for copyrighted code on public code bases. Big tech companies could do the same to squash an open source competitor.

The Ethical problem:

Simply using an AI that's run on a corporate data centers encourages the construction of yet more data centers, with all of the environmental/climate negatives they bring, as well as local harms they induce on the people living near them, such as increased electricity rates.

Using corporate AI directly helps the financial situations of the big tech companies that host them (by boosting usage/user numbers, they are able to attract more investment capital), most of which are owned by right-wing CEO's who are more than willing to collaborate with and fund fascist governments to ensure that they are not regulated in search of both maximum profits. Some of these companies, such as Nvidia, Palantir and Oracle, genuinely appear to be seeking to use these tools for what would previously be considered crackpot conspiracy theory levels of public control and surveillance.

But if someone is absolutely determined to use them regardless, I would hope they either use a locally run LLM, or use a distributed model like AI Horde, as at the very least that does not encourage the construction of corporate data centers nor boost their financial prospects.

4

ProdigalFrog @slrpnk.net 1mo ago

LXQt and MATE are the most comparable alternatives. Neither have commented on AI contributions (either for or against), AFAIK.

ProdigalFrog @slrpnk.net 1mo ago

Using big tech corporate AI in any capacity is a bad idea.

Using a self hosted model to check for bugs (as long as you have the skills to know when the bugs it detects are not hallucinated) is probably okay. I personally would not recommend letting any LLM do the actual coding for you, however.

1

ProdigalFrog @slrpnk.net 1mo ago

Both KDE and Gnome have significantly more developers on hand to contribute toward implementing their own wayland compositors, and on recent versions, their wayland implementations have worked very well for me, even on Nvidia hardware, which I couldn't say just a few years ago.

In comparison; Cinnamon's Wayland backend is only just now becoming usable for most average tasks, Xfce's effort is the focus of this post, which has only just gotten to an alpha state (still not ready at all for an average user or for distros to adopt as the default), and MATE has no developers on the project who have the necessary skills to make a Wayland compositor, so they are instead trying to integrate the Wayfire Wayland Window Manager instead, and they still consider their own implementation experimental.

ProdigalFrog @slrpnk.net 1mo ago

AFAIK, they are wisely not accepting AI code contributions over licensing concerns.

The issue is the lack of software support overall. It's possible to use BSD, but quite limiting.

Haiku OS also decided against accepting AI code for the same reasoning (though that is even more limited, and doesn't support a wide range of hardware, I believe).

ProdigalFrog @slrpnk.net 1mo ago

It's definitely worth buying used wherever possible, and to repair electronics whenever possible to avoid unnecessary consumption (part of why Right to Repair is so important). Similarly, it's quite beneficial to switch to plant-based meats to avoid both unnecessary animal suffering and the carbon emissions and resource wastage that goes with industrial animal production (especially as those resources become more and more scarce due to poor crop yields and drought from climate change).

It's often not possible to do either of those things 100%, such as it not being possible to be completely vegan/vegetarian due to a family member that isn't on board with it, but it's still beneficial to do what they can, even if it's only 20, 30, or 50%.

Corporate AI usage, on the other hand, is one of those things that truly is completely avoidable with virtually no effort, and was something we all got along fine without just a few years ago. There are even more ethical self-hosted or decentralized non-corporate alternatives.

Someone deciding to ignore the harmful externalities of corporate AI is making a choice to prioritize their own convenience over helping prevent our planet from becoming extremely unpleasant for billions of people and animals to live on, or the financing of a fascist surveillance state.

ProdigalFrog @slrpnk.net 1mo ago

It's all about DesktopPal 97 for me!

(Unironically do love that theme, but more seriously, I use Xfce on a particularly weak laptop with only 4gb of RAM, where it works as a good compromise between being lightweight yet not needing the configuration that Window Managers need if they don't come pre setup from the distro itself)

1

ProdigalFrog @slrpnk.net 1mo ago

Of course we can, but that’s not how history works.

History has shown that it does, in fact, work like that.

6

ProdigalFrog @slrpnk.net 1mo ago

Bear in mind that each desktop environment has its own implementation of Wayland, meaning each can offer vastly different user experiences. Gnome and KDE currently have the most polished and integrated Wayland compositors, and are the least likely to have any noticeable issues compared to X11, except for perhaps a few edge cases that have yet to be worked on.

Cinnamon, MATE, and Xfce all still use experimental/buggy Wayland compositors, so if any of those desktop environments were the ones you experienced issues with Wayland on, I wouldn't let that color your expectations of Wayland itself.

2

ProdigalFrog @slrpnk.net 1mo ago

I think it's quite clear I'm against AI use (especially in FLOSS projects) for the reasons I've laid out in the comments here, I make no bones about that. And you yourself are clearly pro-AI, and dislike arguments advocating against its use. You also appear to dislike the idea of individuals being able to contribute toward any sort of societal change with personal choices, instead much preferring to put the onus on a clearly corrupt government which is itself at the beck and call of the very AI companies we're discussing.

4

ProdigalFrog @slrpnk.net 1mo ago

The developer made this public post of their own volition, presumably for others to read and make up their own mind on. It's not naming and shaming to alert others their DE is going to be AI coded in the future, or to point out the potential downfalls of the project adopting AI.

3

ProdigalFrog @slrpnk.net 1mo ago

All of us can try to take steps to reduce our fossil fuel use, either by trying to drive less, using public transport if available, or using an e-bike if viable. Individually the contribution is small, but collectively it can add up to help reduce demand, thus reducing the financial incentive to invest in oil or drill for more.

8

ProdigalFrog @slrpnk.net 1mo ago

Refusing to use the tools that profit the very thing we want to stop is part of standing up to it, how is that not obvious?