All true - but very much written from the perspective of western academia. At the risk of being too condescendingly lenient towards cheats in developing countries, the large body of fake science coming from these (China, Egypt, Pakistan, etc...) is likely driven more by financial need than by personality defects. The enthusiasm in these countries for bad metrics to award jobs (number of papers published, impact-factor fixation, etc...) is a perverse incentive that is likely impossible to ignore when the other choice is eating dirt.
in major ai conferences they have a new rule called "only 10 paper submissions per prospective attendee" and everyone is using this GPTacademic https://github.com/binary-husky/gpt_academic/ (and it's not bad at all at what it does) now we can fraud programmatically , it wouldnt take an hour to "connect it" to submission systems and not even have to worry about deadlines
More interesting question is why write a tweet that is confessing not understanding something that one is in the position to understand perfectly.
Welcome to #Ouchtown!
I would love to watch a synchronous discussion between Heathers and Van Bavel.
I'll let you decide who would play Reish Lakish and who Yochanan!
https://youtu.be/5NRZgRdayPY?si=EWTdYJKtNcjclXrt
All true - but very much written from the perspective of western academia. At the risk of being too condescendingly lenient towards cheats in developing countries, the large body of fake science coming from these (China, Egypt, Pakistan, etc...) is likely driven more by financial need than by personality defects. The enthusiasm in these countries for bad metrics to award jobs (number of papers published, impact-factor fixation, etc...) is a perverse incentive that is likely impossible to ignore when the other choice is eating dirt.
in major ai conferences they have a new rule called "only 10 paper submissions per prospective attendee" and everyone is using this GPTacademic https://github.com/binary-husky/gpt_academic/ (and it's not bad at all at what it does) now we can fraud programmatically , it wouldnt take an hour to "connect it" to submission systems and not even have to worry about deadlines