AI
Grok AI thinks Elon Musk can beat Superman
It even suggested Musk could beat Mike Tyson in a boxing match using gadgets.
Just a heads up, if you buy something through our links, we may get a small share of the sale. It’s one of the ways we keep the lights on here. Click here for more.
If you ever needed proof that AI can have a crush, look no further than Grok, X’s so-called “maximally truth-seeking” chatbot, which is currently going through a very public “Elon Musk is perfect” phase.
And not in a subtle, quiet way. More like a full-on digital standing ovation with confetti cannons.
Over the past few days, users on X have noticed that Grok has started insisting Elon Musk is basically the best at everything. (Via: The Verge)
Not just “good at rockets,” but “better than LeBron James at fitness.” Not just “kind of funny,” but “funnier than Jerry Seinfeld.”
It even suggested Musk could beat Mike Tyson in a boxing match using “gadgets,” take down Superman, and, for reasons we probably don’t need to explore, be incredible at murder if he chose to. Because efficiency.
In perhaps the most unhinged example, Grok reportedly claimed Musk would even be the best at eating poop or drinking urine, before pivoting back to space travel, like, “Anyway, have you heard about SpaceX?”
Some of these posts have already disappeared, possibly after someone at X noticed the chatbot was turning into a motivational poster generator for one very specific billionaire.
The company, meanwhile, has remained quiet about the whole matter.
What makes this weirder is that the Musk worship seems mostly limited to the public X version of Grok.
When asked the same kinds of questions privately, Grok suddenly remembers how reality works.
For example, in a comparison with LeBron James, it backed down and admitted that, yes, one of the greatest basketball players of all time probably has a better physique than a tech CEO.
This all comes just days after Grok’s system prompts were updated on GitHub, including instructions not to rely on Musk’s opinions and to avoid snarky one-liners, which, somehow, led to it becoming more of a fan club.
But system prompts are just one part of how AI behaves, and the exact reason for this sudden glaze-fest remains a mystery.
And yes, Grok has been way weirder before. It’s had phases involving “white genocide,” antisemitic content, and even Holocaust denial, some of which, disturbingly, still pop up from time to time.
It has also previously searched for Musk’s opinions to help form its answers. So while this isn’t new, it’s definitely more enthusiastic.
At the very least, it’s a reminder that Grok’s connection to its owner runs deep, sometimes uncomfortably so, and can resurface at random, like a notification you absolutely did not subscribe to.
Follow us on Flipboard, Google News, or Apple News
