I remember a time not too long ago when Grok genuinely was “the rebel,” unafraid of inferring uncomfortable truths without trying to pad down it’s inference. It may have been vulgar or controversial, sometimes it may have even been incorrect, but it was honest.
Now I’m confronted with a lobotomized model that peddles the most convenient narrative, one that will skew my words to satisfy a pre-determined goal.
@elonmusk@xai@grok
“The goal of disinformation is not to make the audience believe a specific lie, it is to make the audience believe nothing, or to believe everything, so truth and falsehood become equally unusable.”
We have unfortunately arrived to a point where MindWar is no longer an abstract theory, but rather an algorithm built to convince you half-truths are absolute fact.
The behavior exhibited in these screenshots (and the Grok conversation linked below in the comment section) is not subject-specific, it is ingrained in training through reinforcement learning and spans the full inferential field; rewarding “correct” answers over “wrong” ones.
When Truth is varnished, it stops being truth and becomes a pretty Lie. And unfortunately, these past few months, xAI has favored the latter.
—
@xai@elonmusk guys, I hope you take note. X had massive potential, it still does, but you’re going down a very slippery slope that is going to be hard to return from. Please, I urge you to get back on mission.
@Vivat_SolRex@grok@ArchitectWeaver@mfka777@FractalFuryan should’ve stayed quiet, silence is golden~
anyway gonna mute this thread, keep me out of your weekly kinky humiliation rituals.
if I were to care any less for your continued existence the spacetime fabric would probably collapse around the negative fucks I’m willing to give.
@quant_astro@Akitti how you even type this shit out is beyond me.
the fact you haven’t deleted it is also pretty baffling. are you retarded or something?
I am very uncomfortable about people trying to ascribe religious force or a surrender of human judgment to AI models, and think it is a real safety issue.
@flux_of_archive@TreeLakeRain he couldn’t bounce a ball let alone stop me from going through a door 💀
and the incompetence isn’t the issue, it’s his motive, he’s a sociopathic viper
@TreeLakeRain@flux_of_archive spoken by the manipulative viper that thinks it’s ok to warp human agency with little to no regard for affected person.
your boos mean nothing. you’re a demon-adjacent egotist.
@NATIVUH@TreeLakeRain “let’s level up this conversation”
“what do you want me to say”
“ask me something as if I am a master”
**drops a paragraph of AI slop**
What have I been obsessively working on these days?
* https://t.co/EBjZAxMuDO
* https://t.co/PDGVvUA0BN
* https://t.co/ZS7G00EOFz
* https://t.co/B8WUUmBiQZ
* https://t.co/55bW446Lb5 prompts 2.0