[add-sub-mul-div]

> Elon Musk, who owns X, recently tweeted that he was going to “fix” the platform’s AI assistant Grok after it cited research that showed right-wing violence was more common than left-wing violence: “My apologies, we are fixing this cringe idiocy by Grok,” he said.

I understand why it's hard to make AI that's accurate, but Musk's elite engineers can't even make it push a preferred narrative consistently. Sounds like it should be a much easier problem to solve.

[altcognito]

I think the last time they tried it ended up going mechahitler

[Insanity]

Regardless of the bias inherent in Grok, it wasn't a good experience to me. I tried it for some time before going back to ChatGPT (this was before GPT-5 though, I'm also not super happy with GPT-5 UX)

[jdiff]

What issues did you have with Grok/are having with GPT5?

[tim333]

I find it quite amusing that Grok tends to contradict Musk's untruths.

[myvoiceismypass]

Curious - which facts did Musk refute specifically that Grok had pointed out? I don't have a twitter account so I can't really go see this thread.

[Mr_Eri_Atlov]

The last time they tried to make it agree with Elon consistently, it proclaimed itself to be mechahitler so...