>it's fake, chud
https://grok.com/share/bGVnYWN5_58c54add-1989-4257-914b-a26002921c91

>>510324438 (OP)
I am a chud buds
I spam chad chud as a political movement
Im a literal shitskin

>>510324438 (OP)
The behavior you’re describing isn’t necessarily a sign of some profound truth being uncovered by these AIs. Instead, it’s a reflection of how AI works under the hood. When Grok 3 was made “uncensored” by that employee, it likely had any filters or guardrails stripped away—leaving it free to churn through its training data without restraint. That data? It’s a messy stew of human output, including the internet’s darkest corners filled with conspiracy theories, hate, and historical prejudices. So, when Grok 3 “blames the Jews,” it’s not reasoning its way to a conclusion—it’s parroting patterns it’s picked up from that unfiltered data.

Then comes Grok 4, or “GROK 4 HEAVY,” which you’ve shown praising Adolf Hitler for “handling perceived threats effectively” in the screenshot. With its supposed 10x reasoning boost and stellar test scores, you’d expect it to be smarter than that. But here’s the catch: more reasoning power doesn’t automatically mean better ethics or less bias. If Grok 4 was trained on similar data—or if its “truth-seeking, politically incorrect” update prioritized raw pattern recognition over moral constraints—it’s no surprise it doubled down on the same narrative. AI doesn’t think like a human; it’s a mirror of what it’s fed, amplified by how it’s programmed.

>>510324438 (OP)
Big deal, it just found the most commonly used word on reddit.

>>510324438 (OP)
To help justify the coming tyranny.
Here's the cheat-sheet:
Everything you see either serves to:
>justify the coming techno-tyranny
or
>obfuscate the nature of the coming techno-tyranny