I was thinking about why this question bothered me so much.
And I think the reason is twofold:
- The speaker is implying that Grok is some sort of “truth Oracle”, rather than a highly trained stochastic parrot.
- The speaker is implying that anyone, or anything is capable of simply delivering The Truth to them in this manner.
Well, Truth doesn’t work that way!
In most cases, it’s not something that’s just given to you “for free”. It’s something you have to work hard for to obtain, because even when it is given to you for free, you still need to verify it for yourself somehow or otherwise it’s hearsay that you’ve chosen to blindly believe.
And when people imply that these autocomplete robots can give them the Truth, it’s not just wrong but dangerous.
I don’t like it when a bunch of fools are running around trying to turn Grok, or any other LLM, into a Ministry of Truth.
Why aren’t these people asking me, or anyone else, whether something is true? Not that I want you to do that, but note how all of a sudden there are thousands of accounts asking Grok whether something is true.
Could this be some sort of psyop? An attempt to condition the herd to replace their own brain with Grok’s? If so, that can certainly be a step on a road to turning the bot that — contrary to xAI’s advertising, doesn’t actually understand a thing — into Judge, Jury, and potentially even executioner.
We can stop this worrying trend with a simple modification to this question. Instead of deifying corporate LLMs, try something like, “Grok, what do you think?”