@ciaranmak they fixed this one sometime ago, at least in ChatGPT, we need other tests.

@tshirtman @ciaranmak I was wondering the same thing, and verified that someone fixed the answer at least for Google's "AI". Sort of.

They fixed the specific instance it was getting mocked for, but not the underlying issue. 😅

@MySideIsHumanity @ciaranmak oh, fun, IIRC, back then when i tried in ChatGPT, it got strawberry wrong but raspberry wrong.

Now you are testing Geminy (google) though, so different bugs. (and people consider it to be behind the competition).

@tshirtman @MySideIsHumanity

gemini, first answering the question wrong by regurgitating the human response which is to often answer the question wrong, then correcting its response when asked to figure it out on its own.

humans: god ai sucks
ai: no u