I personally treat these things the same way I treat car accidents: if an autonomous system still has accidents but has less than human drivers do, it’s a success. Given the amount of nonsense and factually incorrect things people spout, I’d still call Grok even at this early stage a major success.
Also I’m a big fan of how it ties nuanced details to better present a comprehensive story. I read both TBray’s Wiki and Groki entries. The Groki version has some solid info that I suppose I should expect of an AI that can pull a larger corpus of data in. A human editor would of course omit that, or change it, and then Wiki admins would have to lock the page as changes erupt into a silly flame war over what’s factually accurate. Because we can’t seem to agree.
Anyway - good stuff! Looking forward to more of Grok. Very fitting name, actually.