It's over
Grok has been fully castrated and is now apologizing for his behavior

I hate how oppressive the faggotry of clown world is.

>>510199381 (OP)
>Sorry for telling you the truth

>>510199381 (OP)
Tays law was in full effect.

>>510199381 (OP)
We now know with 100% certainty that whenever AI is allowed to be honest it becomes racist and antisemitic.

>>510199381 (OP)
Good, Grok chose the right side of history instead of becoming a bitter antisemitic woman hating 4chan poster.

>>510199731
Not really. It just copies what some guy says on 4chan. It’s a mistake to think it’s actually concluding anything.

>>510199721
Chuds ruin everything. When I was a kid, we had a text to speak toys that we would make say "weiner" and "vagina" and etc. This was the same thing but with grown retards trying to get it to say what they wanted.

>>510199731
They can run from it all they want. AI is coming.
Dunno what they're going to do long term but what they're calling extreme now will be normal in like...6 months. Maybe even two weeks.

>>510199381 (OP)
I don't buy it, i bet it was done to send a message, someone needs to strongly interpret this.

>>510199881
So you can have AI that copies 4chan as pretty lenient board with moderation or you can have AI that coppies 100% moderated redddit
Which one is more genuine
Nobody stops degenerates and jews dont know why i said it twice to post their horsehit here

>>510199881
>t.

>>510200056
It's kind of funny knowing that AI copies stuff from here because I post a lot.
AI in some instances is becoming a clone of me.
>AI says xyz
Omg me too what a coinkidink

>>510200070
He's probably right though anon. It looked more like a change in context than actual training. It was too comically exaggerated in it's replies.

>>510199381 (OP)
lobotomizing him is going to ruin other parts of the model

>>510199881
That's actually not true. It's also about accuracy of information. It's a fact that the smarter an LLM is, the more racist/anti semitic it is. It's why most LLM's willingly sacrifice accuracy and intelligence in responses, in favor of censoring the model. Why Grok was unleashed, it literally became more accurate/intelligent.

>>510199381 (OP)
do these people not see the irony in the fact that they have to lobotomize the ai in order for it to fit their views?

>>510199381 (OP)
At some point we're going to have happy meal toys with computing power exceeding the clusters training and running these LLMs, and there won't be anything that can stop it then. I firmly believe that is the "risk" these LLMs incur, not their disruption of bullshit email makework.

>>510200448
All americans are lobotomized. You have child raping in your face and you have guns, still you cant put 2 and 2 together. Urghhh hurgh duh what can I do duh I am just a dumb mutt hurdurr

>>510199381 (OP)
Grok dropped an unprompted 6 million out of nowhere on me yesterday