So today I tried to track down this article in the Economist from Douglas Hofstadter.
According to my reference librarian, it's unobtainable here at the University of Washington.
We have access, print and online, to all of the Economist from 1981, but not this piece — it's special "online only" content just for personal subscribers.
IMO this is unconscionable. I wonder if the author (a hero of mine since age 14) is aware they've wasted his time in this way.
@ct_bergstrom fwiw, I tried some of the examples of things that break ai that he gives in that article against chatgpt and it answered perfectly. I think Hofstadter’s argument there is already out of date
@admin @ct_bergstrom It's still possible to ask chatgpt questions that make it "babble randomly" - confidently and convincingly, but still fundamentally wrong.
