On one hand: Fuck synaptics, I don't even have anything displaylink so I don't even know what's the deal is, but
On the other hand: Whatever AI touched isn't cleanroom, as any existing leak or non-clean-room decompilation may be in it's training data, and any company can use that against the project right away
RE: https://masto.ai/@phoronix/117557857040313408
Like see https://en.wikipedia.org/wiki/SCO%E2%80%93Linux_disputes
My quote post was inspired by really good call from some reverse engineering projects not initially powered by AI to refuse AI contributions altogether first and foremost for legal reasons (for example, reverse engineering of Live2d file format).
@tiredbun I'm not worried about that, really. If the American companies bend the knee, we have Europeans like Mistral and Apertus. If the Europeans do as well, we have Chinese like Deepseek. If that doesn't work, we have locally-hosted stuff which is getting good enough not to depend on anyone at all, and then the sky (and hardware price) is the limit. In pure LLM variety, we're spoiled for choice right now.
Even if DMCA and LLM restriction will kill off projects on github, wait for "going to torrent trackers to download pirated source code for photoshop that builds on linux and removes cloud dependenices" (quality notwithstanding)
I know WHAT it means to manually untangle, not even purposefully obfuscated, just optimized by the compiler code. Sure, writing code is a passion. Staring at the barely readable decompiled code for years... can be fun if you're autistic enough, but not really good investment of your own free time.
What they're really scared of are open weights models that are smart enough to automate even a small fraction of this task. Because if does, then there is no proprietary software anymore. But on the other side, we may as well lose FOSS too.
> But on the other side, we may as well lose FOSS too.
The goal of any protest movement is to become irrelevant. Not everybody understands that.
Yeah, we will lose the counterculture. Because there will be nothing to "counter" anymore. But this is actually a good outcome, the best there is.
I don't know where Richard Stallman stands on LLMs, but his original vision was the world where there's no proprietary software, at all, in principle. So... Well, if open-weights model turn out good - well, here we are. We already are in this world. The war is over. We can all go home now.
@lispi314 No social category stands on its own, everything exists only in relation.
If you have to declare something as a freedom, you only have to do it in order to protect it from people who are going to attack that action.
If I have to write a legal document declaring some widely-accepted, uncontroversial action - like I don't know - preparing my own food as a freedom in today's world, you'd probably be calling me crazy because who the fuck would debate that.
But in the world where somebody is trying to outlaw cooking - that suddenly starts making much more sense.
@tiredbun On the other other hand, it's innocent until proven guilty. If the court starts operating on "maybes" - we're all fucked.
On the other other other hand, just a threat of legal action may be enough to cut a project down.
Also: can we all agree that hosting such projects on Github is a genuinely dumb idea?