lag those of an average mathematics graduate student, according to a paper published on Cornell’s Arxiv. About a ten minute read for the paper, abstract is a one minute read.
Comment: it’s pretty clear ChatGPT gets basic calculations wrong, but I’ve found any math problems requiring deductive reasoning totally confuses the bot. SO as educators we need to insure we’re not asking simple ‘solve for Z given X and Y’ type questions, rather given X what characteristics of Y would result in Z’? More thinking, even better cross domains, and more critical inquiry is needed.
