you don't really even need to switch AI engines, just feed the same response into different chatgpt sessions and it'll eventually answer wrong and tell you that it didn't make it , anyway.

this whole truthiness thing surrounding AI outputs is really the most entertaining concept of this whole trend for me. I wouldn't have guessed that one of humanities biggest teething-points w.r.t. AI is vetting that the confidently given answers were even correct.

I don't know if it's irony or not, but there is definitely some humor in producing a confident liar when trying to solve the problem of AGI.