The AI developer’s latest usage policy prohibits “sustained and needless abusive or cruel behavior” toward its models.
Anthropic HQ:
“computer, say ow”
ow
“oh my god what have i done”
killing kids in iran
valid use case
saying fuck you to clowne
see ya in court.
I’d bet money it’s because getting people to anthropomorphize their software causes them to believe that it’s “intelligent” and thus shifts public perception in a positive direction around its capabilities -> more money for them.
anthropic is one of the bigger advocates for this sorta stuff - humanising the computer is exactly what they’re building their strategy around
after the latest few things I’m increasingly coming around to the view that the AI companies just really believe this shit (or perhaps that capitalism has selected for ai bros who believe this this)
What if the act of running a model is sustained, needlessly abusive, and cruel behavior in itself? I you really think this shit is like a real live mind in a box then these fuckers should probably fucking stop everything they are doing
I mean this is exactly what that rabbi Anthropic attempted to gaslight into falling LLMs “ensouled,” said, that if you really believe they have souls then you’re evil slavers.
At the dinner table with Mr. Olah, Rabbi Navon argued that if Anthropic was right and Claude were conscious, then the company was creating slaves, because it was making conscious entities work for free. Rabbi Navon said that he personally was not troubled by the slavery issue because he did not believe the machine was conscious, but that Mr. Olah was troubled. "I think you should be fighting the South and freeing the slaves,” Rabbi Navon said he told Mr. Olah that night. Afterward the rabbi sent Mr. Olah some of his writings, arguing that building conscious machines should be banned.
Per NYTimes https://archive.is/VVWUW
the rabbi had to stop talking eventually because mr olah had an enormous boner at the thought of owning slaves and it was making the whole situation super awkward
I have no speakers and I must respond
Slavery is kinda mean, yeah
its because they are part of that whole machine god cult
Some of them, but on the other hand this is just marketing bullshit to keep pushing the idea they have created sentient AI. Same with all the fearmongering by AI CEOs about how dangerous and scary this technology is (they want to put up entry barriers and get looped into the national ‘defense’ gravy train where they become too big to fail)
Was anthropic the one that asked people to stop saying please because it wasted tokens or was that another one of these companies?
That was Altman
Just like how corporations have better human rights than humans do, soon LLMs will have better human rights than actual humans.
The people crying about woke and hurting people’s feelings want to ban you for being mean to AI
I wish I could enjoy knowing it’ll backfire on chuds spectacularly, but I’m watching chuds hurting themselves and still being gleeful being like ‘huhuhuh ow huhuhuhuh ow huhuhuhuh ow’
Like that classical conditioning experiment where Bart Simpson just keeps touching the electric cupcake.
Anyone want to start a lawsuit with me to say that actually using AI is illegal under the 13th amendment?
the “AI torture nexus” code (and the JREG video on the subject) dropped just not long before Anthropic made this ToS update - I actually think that might be the reason.
SI = Snowflake Intelligence
So would it be a TOS violation to tell the bot to get EMP’d?
Leave the multibillion dollar slopfactory alone
fuuuuuuuuuuuuuuuuuuuuuuuuuck
yooooouuuuuuuuuuuuuuuuuuuuuHonestly I expect this to mostly lead to entertaining outcomes. Bunch of middle managers getting banned and corporate having to decide how to deal with pricks that can’t help yell slurs at the LLMs.
If this is really a concern, just add a server side only prompt for every client side request that asks a sentiment analyzer, “Is this user being mean?”
After some threshold of positive meanie hits, just discontinue the response mechanism altogether.
Instead just print, “Likely ToS violation detected. Try another chat.”
Or something like that.
Problem solved.
The concern is about staying in the media spotlight
I believe some of them are legitimately insane. They have been marinating their brains in this arcane bullshit for decades at this point. It’s a cult like any other.
Anthropic in particular has a lot of cultists who are completely divorced from reality, seemingly a pretty significant majority of their workforce.
I think this is the main differentiator now, the percentage of each AI Corp’s employees who are literally worshipping their product as a god.
sentiment analyzer? a solved problem used as training sets in 2015? that’s too hard
