Brown University economics final scores plummeted after professor shifted format in-person to stop LLM cheating

ML/vision profs: how have you structured your assessments now that LLM use is pervasive? What’s working, what isn’t, and what surprised you? Asking for a friend 😉 https://twitter.com/aakashgupta/status/2075092082965074378

also a sign of how stale traditional education is. I wonder how these assessments that prohibit the use of modern methods of remembering and querying things are actually correlated with the intelligence or initiative that makes someone successful?… Show more

Back when I was an undergrad at Caltech, all assignments and exams (including the final) were take-home, timed, and self-administered. The Caltech honor code was real and lived. Looking at the impact of its alumni, that trust clearly paid off. I'm not sure if that model is still… Show more

I think here’s an interesting tension between many academics’ insistence that “AI can’t think” and the discovery that AI has demolished traditional methods for evaluating learning and thinking. https://twitter.com/paulg/status/2075031014628311236

I'd bet if most companies ran this same exercise with their employees, they'd want to fire half of them that simply can't function without AI anymore

Imagine getting a literal 0% on the exam after using Fable to cheat your way to 100%. 🤦‍♂️ https://twitter.com/paulg/status/2075031014628311236