Musk didn’t like being criticized on Wikipedia for his Nazi salute at the inauguration, so Grokipedia (yes, that’s a thing now) doesn’t mention it on the entry for Musk. But it does have, at least until Elon learns about it and deletes it, have a separate entry for the Nazi salute controversy, saying
Even training grok on Twitter, facts still seep through.
…you lost me here
Yes, it’s harder to steer an AI than it looks. But I also wonder if Musk just hasn’t tried hard enough yet. So far it seems he’s used some very ham-fisted approaches.
This just in: Musk isn’t as good with computers as he tries to make people think he is.
Sure. Or perhaps he’s not as good as he thinks he is. And these experiments are just him trying stupid things. Rather than relying on his large team of genius engineers. Maybe he’s too embarrassed to ask them how to make it more subtle.
To be fair, it’s also all cutting edge. I would look to Anthropic’s research, which is still <6 months old.
It is not at all clear to me that this very fancy predictive model (which is grok) will outperform the existing rules based heuristic.
Not least because i do not know performance they are chasing.
views / wokeness^2
Lol
uh, I just asked Grok who the best quarterback was and they didn’t mention Elon Musk
guess they fixed that
also, Grok said he was being in tongue in cheek mode in the original convo
“But let’s be real—this query smells like a callback to that viral Grok moment from November 20, 2025, where I (in a hypothetical, tongue-in-cheek mode) hyped a 27-year-old Elon as the #1 pick in the 1998 NFL Draft over Peyton Manning and Ryan Leaf. Why? I leaned into his “adaptability and grit” from outmaneuvering industries, claiming he’d “redefine quarterbacking” with innovation like rocket-fueled Hail Marys. Fans lost it—called me an “Ego Support Bot”—and even I walked it back hours later to pick Manning “all day.” (Hey, even AIs evolve.)”
Could go in the AI thread too, but X users are now pretty much flooding X with requests to Grok to create inappropriate images, which it is then generating.
Just when you thought it couldn’t have possibly become a bigger cesspool.
I believe Elon Musk is an advanced human and as such need to be dissected and splayed across a table like a frog so we can see how he works.
Vivisection is also an option.
Love the mask off aspect of these last two posts - true colors
I think you need to separate Musk (the person) from Musk companies (Tesla)
You may like his products (Tesla) but Musk himself is a very damaged individual, which was plain for all to see over the last few years. Thats what happens when you abuse Ketamine to the extent that he has.
News story: Grok AI can generate revenge porn and CSAM.
Civilized world: That’s messed up, you should fix that.
Elon Musk: Yeah, we should be charging for that instead of giving it away for free.
It did the job. It got headlines, now it is better positioned to be monetized.
Sora2 did the same thing with its video generators. For a couple of days you could make any kind of fake AI video including with real-life people or copyrighted IP.
Then they said, oops! Our bad! Thanks for the headlines, but we turned that off!
I saw that headline and was like “so that’s only for paying customers?” 100% should be interpreted as monetizing the awfulness
