This can't possibly be the case, because the AI voice says slightly different things between the two attempts. The first time it says "you've already combined the base ingredients, so now grate a carrot to add to the sauce"; while the second time it says "you've already combined the base ingredients, so now grate the carrot* and gently combine it with the base sauce".

Unless you think they've added some inference logic on the device to slightly re-state the last answer they got from the cloud, it's clear that the glasses were connected and receiving the same useless answer from the cloud.

* side note, but it can also sound like "pear" to me this second time

Oh could very well be the case I've only listened once!