@tokensane
I'm confused. you mean you were pessimistic, right? LLMs can be more effective than you had thought
@tokensane I'm not clear on how the course was degraded if students understand the course material at the end. My question would be on whether the students discriminated between correct and incorrect assertions the LLM constructed when they wrote their essays. It's like, when I use glide typing on my phone keyboard: I can still identify and correct errors while saving some effort in typing. Another analogy I like is to the calculator: if I don't know which calculations to do, or if I input the wrong numbers, my lack of knowledge will still show on examination when I fail to realize my answer doesn't check out.
@2ck I meant that my belief that ChatGPT wasn't already degrading undergraduate courses was optimistic.