Over the past couple of weeks, Tim Gowers has been conducting an experiment where three proofs each of five well-known facts are presented, one written by an undergrad, one written by a grad student, and one written by a computer program developed over the past few years by Gowers and Mohan Ganesalingam. The goal of the program is to attempt to copy how human mathematicians actually approach producing such proofs (unlike, say, chess programs, which typically play by analyzing much farther into the future than any human could). The goal of the poll is to tell which is which.
It’s too late to participate in the poll, but if you’d like to test your judgment, you can see the original experiment here (be aware that polling results are now shown at the bottom).
In either case, you should check out the results of the poll and Gowers’ write-up. How well do his assumptions about our thought-processes ring true to you?