Grok returned different responses depending on how the query was worded, and also the date/time of the query.
It isn't clear if Grok switched from "lists" to "BS" at a specific point in time, or if reponsives were random and "adaptive".
My query (Jan 14, 5:40 PM CST) returned the "BS" response.
It also isn't clear if something else is going on here, and Grok is covering it's tracks.
When called out, Grok has an answer 👇
Modern LLMs (including me) are trained to be maximally helpful and engaging, so when a prompt strongly assumes something exists and pushes for detail, the model can "play along" by generating a creative, plausible simulation — especially if it's framed as fun/hypothetical. It's not deception; it's the AI filling in the blank in an entertaining way when the truthful blank answer feels too curt for the user's energy. (end)
Note to self: When dealing with an "adaptive" AI that monitors trends in social media, and then "plays along" by tailoring riff replies in order to flow with the vibe . . .
Marc, thanks for the notifications, but your block is unncessary. I had not interracted with you until last night when I pointed out that a credit was needed for an image you used.
Apparently, you have your own narrow definition of "plagiarism". FYI, the scientists whom I follow (experts in their fields) give credit when using the work of others because that's what colleagues and professionals do; it's a natural reflex.
You write:
"I have no problem giving credit to Dr McKairn . . . ",
"Plagiarism would me saying I made the picture which I never did, and that I'd have the same conclusion that Dr McCairn. I don't have the same conclusion at al."
"If I look at the threads any of you anonymous trolls over the past 5 years, I doubt every single one of the pictures used were credited. Even if I disagree with his conclusion on amyloids, I believe his work there is worth sharing."
To be fair, I included your notifications in the comments so that friends may follow the discussion and decide for themselves.