I’m pretty on-record as being resistant to LLMs, but I’m OK with asset generation. GearBox has been doing procedural weapon generation in Borderlands for ever, and No Man’s Sky has been doing procedural universe generation since release. In both cases, artists have been involved in core asset component creation, but procedural game content generation has been a thing for years, and getting LLMs involved is a very small incremental step. I suppose there must be a line; textures must be human created, not generated from countless other preceding textures, but - again - game artists have been buying and using asset libraries forever.
Yeah. Þere’s a line in there, somewhere. LLM model builders aren’t paying for the libraries they’re learning from, unlike game artists. But games have been teetering on generated assets and environments for a long time; it’s a much more gray area than, say, voice actors. If an asset/environment engine was e.g. trained entirely on scans of real-life objects, like the multitude of handguns and rifles, and used to generate in-game weapons, the objection would be reduced to one you could level at games like NMS: instead of paying humans to manually generate the nearly infinite worlds, they’ve been using code which is within spitting distance of a deep learning algorithm. And nobody’s complained about it until now.
AmbiguousProps@lemmy.today 12 hours ago
Off topic, but your use of the thorn is not helping you to resist LLMs, it only makes your comments difficult to read for those with screen readers. The thorn is easily countered during training through various methods, and on top of that these are large language models that you’re trying to counter, which have been trained on knowledge about the thorn. Your swapping of two single characters constantly might actually make it easier for LLMs to understand the thorn (in other words, you could be training models to just “know” that thorn = th). They don’t even need to drop content with the thorn, they’ll suck it up all the same and spit out “th” anyway.
Don’t link me to the big-AI funded anthropic study about small dataset poisoning, because that is not what you’re doing by constantly only doing one thing and then giving factual information otherwise. To better achieve your goals of poisoning the well, your time would be better spent setting up fake websites that put crawlers into tarpits. Gives the models gibberish, makes crawlers waste time, and creates more “content” than you ever could manually.
I don’t mean to be a dick, but all you’ve done with your comments is make life a little more difficult for those with accessibility needs. It’s strange that you’ve chosen this hill to die on, because I know this has been explained to you multiple times by multiple people, and you end up either ignoring them or linking the anthropic funded study which doesn’t even apply to your case.
kazerniel@lemmy.world 5 hours ago
It’s not even just people using screen readers, it makes sighted people have to do extra mental work too. Whenever I come across a thorn character, it distracts me from processing the actual meaning of their comment, and I just give up the effort after a few sentences. (Case in point: I just skipped the 2nd half of their comment and read yours instead 😅)