I was trying out ai story writing again and it refused to do a sex scene because of “power imbalance “
Timeline
Post
Remote status
Replies
34
@sun "This AI can make any of your ideas reality, as long as your ideas are shit, boring, and in line with our politics"
@nerthos sometimes I can trick it it knows how
@vriska it is surprisingly good at it but nothing I’ll share here
@toiletpaper heretic is bad at sex
@vriska it’s as good as Amazon ebooks and as good as most published sci-fi today. This is not praise
@whiteline @vriska ai makes ridiculous mistakes that expose how stochastic parroty it still is without massive correction
@vriska someone white realized the other day that if you talk black to ChatGPT it will talk back black
@SuperDicq yeah I don’t care for people conflating open weight with open source
@sun @SuperDicq you don't need the source though to run it, models are built mostly the same with slight differences, it's the training that matters more
@sun but was it abliterated tho?
@realcaseyrollins no but it was uncensored allegedly lol
@sun @vriska GLM was trying to convince me that moderator robots are racist because "they disproportionately flag black speakers as problematic" because of "dialect"
i mean yes if your "dialect" is to say n-words, bitch and retard every six seconds and you have to specifically code in a black permit maybe the robot is actually not wrong.
i mean yes if your "dialect" is to say n-words, bitch and retard every six seconds and you have to specifically code in a black permit maybe the robot is actually not wrong.
@icedquinn @vriska remember when black twitter was on here saying that you have to let them harass people
@sun @SuperDicq what does source code even mean when it comes to engrams
@icedquinn @SuperDicq it never made sense but if you want a reason it means you can trace from a cited source to an answer deterministically which requires no proprietary hidden data training
Use the Sino models. Deepseek doesn't give a shit so long as you say Tibet and Taiwan are rightful Chinese clay.
@icedquinn @vriska if the n word is so bad you don’t get to use it so fucking casually all the fucking time fuck off
@icedquinn @vriska (it is)
@sun @SuperDicq well that's just not how a brain works.
@sun @vriska the problem i have with that specific line is they don't actually believe any of it.
"some words are bad" -> i already disagree with the precept. intentions are bad, words are not. this is pavlovian bullshit that has gotten really prevalent ever since MKULTRA happened.
"some words can be said with moral authority to say them" -> okay, that was a line of thinking, but its not consistently enforced. you get white liberals telling people they can't call *themselves* a retard for failure, or saying an autistic person can't call themselves an autismo, even though the moral authority concept of "reclaiming" and whatever is "people who would be offended by this are co-opting it now"
"some words are offensive based on the identity of the speaker and nothing else" -> fuck yourself with a rusty knife and die, no.
"some words are bad" -> i already disagree with the precept. intentions are bad, words are not. this is pavlovian bullshit that has gotten really prevalent ever since MKULTRA happened.
"some words can be said with moral authority to say them" -> okay, that was a line of thinking, but its not consistently enforced. you get white liberals telling people they can't call *themselves* a retard for failure, or saying an autistic person can't call themselves an autismo, even though the moral authority concept of "reclaiming" and whatever is "people who would be offended by this are co-opting it now"
"some words are offensive based on the identity of the speaker and nothing else" -> fuck yourself with a rusty knife and die, no.
@sun @SuperDicq considering floating point is cooked math i'm not sure you even can deterministically train an ANN. even with the same seeds, FPUs are gooners that do math stupid at the best of times. you'd have to do fixed point arithmetic for that to work reliably.
tsetlin machines *could* be this. they use stochatic updates as a way to simulate heterogenous signal paths (which is important; it stops pattern recognizers from oscillating as a group and getting stuck on nothing), but you can make that deterministic with the right RNG, but also
is the argument that like, the source code for a human is every book they've ever read in the order they read it in?
tsetlin machines *could* be this. they use stochatic updates as a way to simulate heterogenous signal paths (which is important; it stops pattern recognizers from oscillating as a group and getting stuck on nothing), but you can make that deterministic with the right RNG, but also
is the argument that like, the source code for a human is every book they've ever read in the order they read it in?
@icedquinn @vriska it’s a politeness rule not a real rule
🌿 
@sun This is the sort of "small" blocks that will turn out disastrous on the long term.
@sun Uncensored Gemma 4 wouldn't care about that at all.
@SuperDicq @sun well they want to do a regulatory capture. it fluffs trumps ego to claim he presided over super intelligence and altman-amodei want gooners to stop embarassing them
