> On two occasions I have been asked, 'Pray, Mr. Babbage, if you put into the machine wrong figures, will the right answers come out?' I am not able rightly to apprehend the kind of confusion of ideas that could provoke such a question.
Impeccable quote. I suppose the interesting thought experiment here is, what if Babbage is wrong? I don't know the answer here, but (and go with the thought experiment) what if model collapse wasn't an inevitable outcome of feeding the snake its own tail.
It can be. The nice thing about AI is that it can create much faster than humans. The problem with AI is that it can create wrong information much faster than humans. This can pollute the sources of information for future AI.
Also consider: "previously correct" is the same as wrong.
