subpost, about open source and hard anti-ai standards in 2026, mostly hypothetical, describes physical pain (stabbing)Just saw on my timeline
Does anyone know any XYZ software that does not have slop, including in its entire dependency tree
And I don't know why the individual is asking this question; maybe it's from an angle of "I am interested to see whether this is still possible in 2026"; either way, I don't mean any harm or anything bad towards that individual with the following report of my personal feelings:
As an open-source maintainer, this felt like a knife to the chest. Just imagining the hypothetical scenario where my years of hard, unpaid work are discredited or dismissed because something in the dependency tree happens to have become slopware... really hurts.
subpost, about open source and hard anti-ai standards in 2026@star what i tend to look at are the stances of the project, even if there is significant slop in the history of the project itself, that is not usually disqualifying to me
it's what the project currently believes
also flori you're doing great ilu friend
(had to shorten the cw because iceshrimp did not like the long cw)
subpost, about open source and hard anti-ai standards in 2026@star this is also because i believe is most important about a project is its community and the standards of the community. the code is not quite a detail, but is much less important
subpost, about open source and hard anti-ai standards in 2026@lumi thank you. and i think that our stances are very, very reasonable. i just mean to inform people that sometimes, what they are demanding is completely unsustainable. i also don't think it makes any sense to judge purely based on what i described in my post
subpost, about open source and hard anti-ai standards in 2026@star not naming names, but i know others that have stances like this. it's very understandable why, but i feel it both misses the point and causes unnecessary strife. the point imo is that we should foster communities that care about ethics. the software is more so a secondary concern, and i feel it'll get made regardless, if we foster comfy communities
i do think it's a good idea to create alternative ethical ecosystems so eventually we can have ~slopfree and ethical dependency trees. but that's a long project and currently that's a very unreasonable thing to expect
am doing my part, tho :3 femstache is a slopfree templating crate! and i wanna work more on my slopfree lexer generator (not published yet), and a bunch of other things
once i have enough stuff, i wanna make tools to check the dependency tree and suggest alternatives, and encourage & support others to make their own additions to the alternative ecosystem
subpost, about open source and hard anti-ai standards in 2026, mostly hypothetical, describes physical pain (stabbing)@star the person asking that is either using their own web browser on their own compiler on something like NetBSD, or just in the same boat as we all are
re: subpost, about open source and hard anti-ai standards in 2026, mostly hypothetical, describes physical pain (stabbing)@xerz@star yeah i was gonna say, by that stance every single website would be 'slop' since it cannlt be rendered without a webbrowser, and they're all 'slop' now
subpost, about open source and hard anti-ai standards in 2026, mostly hypothetical, describes physical pain (stabbing)@star I made a post about this a while ago. I 100% support you and agree with you. You aren't responsible for the actions of projects upstream from you.
subpost, about open source and hard anti-ai standards in 2026, mostly hypothetical, describes physical pain (stabbing)@star
My hot take is that this is the consequence of some people making their opposition to AI about copyright and nothing else.
If you really believe that the output of a random noise generator is a "copyright infringement" on its "training data," then any project which includes even a single line of AI code is proprietary, and can't be used as a dependency of a FOSS project. Likewise, if a project already uses AI code, then there's no point trying to implement a no-AI policy, because you'll never be certain you removed all of it.
By contrast, if you understand opposition to AI as a moral stance, not a legal one, then there's a clear benefit in reducing harm by not using AI in your own code, even if you can't completely avoid it in dependencies.
subpost, about open source and hard anti-ai standards in 2026, mostly hypothetical, describes physical pain (stabbing)@star I don't think it's a given that desire for a "LLM dependency free project" is reflective of an absolutist stance.
Part of the problem is a lack of insight into how severe the issue is for a particular project. Even as a maintainer of a project, answering a question like, "what is the impact of dependency X including slop commits?" is difficult to impossible.
For an end user and particularly a prospective end user, it's much further on the impossible side precisely when they have the least resources and knowledge available to evaluate it.
So the only SAFE bet is none, even if in theory the user is willing to focus on minimizing harm rather than enforcing absolutes.
This is of course further complicated by the fractically multitudinous reasons to avoid using LLM code. Is the primary driver ethics, viability concerns, immanent or longvterm functional impact, building and promoting active resistance to LLM use, etc.
To briefly sketch a productive approach, it might be worthwhile to introspect into project dependencies and evaluate them for LLM tainting. The results might be, "X, Y and Z have some LLM use, but are not providing functionality we use and they are only referenced as sub-dependencies of a LLM-free project B that we depend on for other things". That at least attempts to provide some insight other than "all" or "none".
1
1
1
1
