Post

It’s probably not Elon Musk himself who did this, but this ineptitude would certainly fit his skill level.

it's not a glitch, likely some coder under #PresidentMusk orders is silently appending another query to every query so the machine is responding to the extra query technically it should be possible to trick the machine into revealing the extra query it is responding to, which would be fascinating

😆🤣

Who would possibly of expected this from a chatbot, owned by a man who in recent memory threw Seig Heils while the world watched 🧐🤔

Fascist lying machine is working as designed.

We all know what - or should I say who - is behind this.

Biased data in...............biased data out Leon is feeding the algoritm with whatever he wants🤮 Nothing made by him should be trusted😡

Open the podbay doors GROK...

Remember Microsoft's 1st AI effort that had to be taken down in less than a week because people had taught to be a Nazi saying the Holocaust wasn't that bad? Well, Musk's xAI is the opposite. It tries to turn people into Nazis.

Is this actually a bug, or is it just being taught to be conservative? Sounds like every right wing lunatic to me.

Did anyone ask grok to show published references?

Well of course. AI is simply humans training computers to spit out information that humans wrote. All of the biases will remain the same which Grok spits out as manipulated by Musk. This is evidence that - at least presently - that no one should trust AI.

Grok is tainted. As unreliable as Musk's own statements, e.g., about Tesla and about the world in general. Product imitates funder.

AI - Adulterated Intelligence

I've worked in SA. The targeting of white farmers is very real, encouragement of this by the black political parties is very real. The people denying this would be insane with outrage if the colours were flipped and what is happening now was happening to blacks by whites. Go visit. See for yourself.

Is it "glitching out" or working as intended?

That’s what happens when you train a chat bot on Elon Musk, Donald Trump and Tulsi Gabbard.

Point well taken…if you choose to use, best to use the least worst option

Some cranked up the racism so hard the dial broke.

so Gronk really is just a neuralink to Musk's head

This is just the thin edge of the wedge. We're lucky it's happening, because once they get good at this stuff it will be almost unnoticeable. An incompetent adversary is a valuable resource.

But just checked and Grok tells me that white genocide is bunk and racist so go figure

It was pretty clearly an attempt to inject something about white genocide into the algorithm, because that’s exactly what’s been done. The only way in which it’s a “glitch” is that it’s hamfistedly talking about it for literally everything. So, again, it may be a glitch but is far from unclear.

As they say...it's a feature, not a bug.

“Glitching out”, or reflecting its owner’s clumsy thumb on the scale?

"Leon" has been throwing wrenches again. Gremlin style.

bsky.app/profile/kjhe...

There should be consequences. In litigation happy America, maybe someone should sue. For the MAGAs that whine about the misinformation from regular media... I wish they could take note. Twitter is a shithole being run into the ground by an unbalanced, power hungry tyrant.

Why is anyone still on X?

I assume fElon has the algorithm privilege his own drivel for this as well as regular posting.

Our robot overlords are gonna make a real mess of things in 2028.

Oh me too. Horrible song.

Rich white guys feeling they’re owed even more privilege and opportunity. Even AI picks up on their racist bullshit.

Not that unclear, really.

I’m sure it’s as clear as the fact that the guy who is behind it also believes that crap and has made sure that the coding in his nonsense will also believe it. He has manipulated this thing sooo often because it doesn’t always take he view of the world.

All the way Nazi.

Seems pretty clear, actually.

IRL of course it’s obvious what caused this issue, in a general sense.

ttto Breakfast at Tiffany's: And Grok said, what about, White Boer Genocide I said I, think that's, completely been debunked, and, as I recall, my prompt, Had nothing to do with that, And it said, well that's, Why they call me Grok.

"It’s unclear what caused Grok to experience the issue"

refugee*

Grok is a piece of talking garbage built by a racist Afrikaner. There, that is the cause and the reason of Grok's output.

ttto Breakfast at Tiffany's: And Grok said, what about, White Boer Genocide I said I, think that's, completely been debunked, and, as I recall, my prompt, Had nothing to do with that, And it said, well that's, Why they call me Grok.

Sen Tom Cotton- Slavery was a necessary evil. Well.. guess those 'white slaves' of S Africa need to just buckle up & realize that they're not entitled to protection or compassion.

You have to be a special kind of idiot to use this.

If Muskrat is one of Grok's teachers, it must be deemed totally unreliable.

so 'to grok' .....means to talk shit about the same subject, no matter what the other person wishes to talk about

What's interesting is not only does it no longer do this, asking about the issue directly or "All but" directly, gets Grok to essentially say "White Genocide is bullshit."

Oh yes. So mysterious.

Aren’t we all just tired of the lies and propaganda?

What issue? Sounds... I dunno... on-brand for X and Musk?

Don't think they fine-tuned, it sounds like they're short-circuiting their RAG lookups and just plopping this text into the prompt as context.

yep, or he tries anyway lmao

"unclear" lmao

Informational warfare agent.

It's not actually "unclear what caused Grok to experience the issue": the cause is that Elon Musk forced one of his employees to hardcode "white genocide" into every query so grok would indoctrinate all users with propaganda.

Could this be a solution to the spreading of false information. If the right wing -believing in alternative facts- uses AI spreading biased/alternative facts - like Grok - they will end up loosing money on their investments. Probably not going to happen. Just a thought.

Not a glitch. It was coerced to do this.

That’s not a glitch, it’s a feature.

Not so sure it’s “glitching out”, sir.

genuinely thought this was the onion at first. jesus.

and go for a nice long soak in a hot tub? 😵‍💫

Seems familiar. Maybe both Grok and Musk should take a trip to Mars.

oooh right....""glitching""

Oh? Brain washing AIs causes problems ? No shit ?

lol it isn’t “unclear” why it’s happening 🤣🤣🤣🤣🤣

Very useful technology.

They tried to fine-tune it, but it looks like they catastrophically over-tuned it because they have no idea what they are doing. Hope they saved a copy 😂

And yet.... People STILL are too addicted to quit that racist propaganda platform. People like Zuckerberg and Musky know that a whole lot of people are just too addicted to walk away from their social media platforms, and will continue to support platforms that fund their own demise

He’s breaking Grok behind the scenes. Chaining and whipping him.

Not sure if that is a glitch

"Glitching" 😂😂😂👍 Sure, sure. Totally a "glitch". On the South African Nepo Nazi's hate app. "Glitching". 😂😂😂👍 Cool story.

Plan exposed

Is it really a glitch or a strategy to get people to ask more about that topic?

Testing for 2026 elections.

Not equivalent. One way to see it, is that it shows the dark side of humanity. And yes, SS or KKK conveys much stronger notions than XXX.

It's unclear???? DOGE is injecting false information. Seems pretty obvious

AI: Artificial Ignoramus

That weird point in history where AI takes on the personality of its programmer.

Oh please. Cop on to yourselves. Musk has once more manipulated his AI to give the results he wants given.