i mean… i’ve seen that kinda porn on furry sites. there’s a whole subgenre. usually the animals can talk, but they’re still animals. as in, not anthropomorphic.

not that i really want to continue the discussion (we agree that SA is bad and the people doing it should be put in prison) but isn’t this a bit like saying furry porn is pretty much the same as bestiality?

not intrinsically, at least not for long. in most of europe it was legal up until the 90s and there were even porno mags that ran specials.

…i do compulsive late night wiki walks, before you ask how i know that. i had a period where i read up on what happened to the “flower power” generation with all their “free love” stuff and that came up. i tend to do breadth-first search so i just read through articles while opening all the links in the background, then go down a level, etc etc. when a term like that comes up on a wiki page it triggers your fight-or-flight response.

i don’t think this is normalising, is it? if anything the tone of the tweet almost makes it feel like they’re offended that the two things are being compared.

oh i am anti as all hell. i have knowledge of how the models tend to work, how the training is done, and how reasoning and tool use is implemented. that’s why i never use online models and check the datasets of local ones.

but like… the idea isn’t bad. it’s just that the way people use it is to steal information, burn through massive resources, and atrophy their knowledge. just like how the idea of a block chain, or a uniquely identifiable digital token, isn’t intrinsically bad, but the biggest use was grifting.

a model trained on consensually acquired data, using only renewable energy, run locally, can still achieve useful things like quickly finding information from a loose query, doing bulk edits, making templates and placeholders, or offering suggestions. it’s just that it would take longer and not be hype-worthy. so no for-profit entity does it. that’s what makes projects like the ai horde interesting.

it’s the curse of getting into something early and it becoming a global issue; i did a course on machine learning in 2014 and have tinkered since then.

like imagine getting into blockchain-based systems because of git in 2007. can’t talk about that stuff at all now.

whenever i see “onii-chan” written out nowadays i mentally translate it as “hey bro”

especially in the context of an in-universe character saying it, yeah

i will always take the time to explain things i find interesting to people. the benefit here is that i can now much more efficiently break large models as well when i come across them. helps me add anti-ai clauses to websites, cv’s, and repos i publish.

yes

yeah, it’s a great way to see the limitations of these systems. just like ctf and ioccc.

dual booting just makes things more difficult. windows likes to silently change things in the bios and on disk drives it manages, basically pulling the rug out from under linux. you can mitigate it by making linux aware of the fact but that’s not the default behaviour, because again dual booting isn’t really recommended.

arma is an annoying one because it’s a bespoke engine with a lot of deep windows integration, coming from the fact that parts of it is from 1995. add online multiplayer anticheat to that and it gets even more difficult.

that said, i got it running on linux in like… 2019, so i would have thought it to be easier now.

i refuse to use ml models for code. the copyright issues alone should be enough to keep them away from every public code base until the matter is settled. but also because local tooling is, frankly, shit. i have a bit of hope for text diffusion models, but i have a hard time seeing the situation improving because everyote is full in on cloud models now.

i do. i experiment with transformer and diffusion models like a few hours a month, tops. the result isn’t interesting enough. the process of bending and breaking the models is the fun part.

helping the separatists prolly

CUTE

i think the first qiestion i have is, what use do you see in this? because… it’s a readme file. it’s written once, maybe updated when the build process changes, but it’s generally static. what’s more, it has no required sections, many projects don’t need examples, and most things on the big forges don’t have links. it’s just a freeform document. it doesn’t even have to be markdown. rst is pretty common as well, as is plaintext. are you suggesting a tool that can parse a file of unknown format and figure out whether it contains predefined structures?

you know, i’ve tried to defend some usage in the past, explaining my processes and the many steps of manual refinement, masking, and layerwork i put in to things, how i only run local models with open weights, how all my power comes from hydro etc etc

but as the tools keep evolving i’ve realised nobody else seems to actually care about the process. the pro-people just want as much slop as possible. someone likened it to a slot machine, where you keep pulling just because. that’s where we are now.

it’s not. demand just dropped off a cliff.