Yumechi | ゆめち   Nov 21
"I set a trap to catch instructors reusing 2010 slides, the results were shocking."

Academic integrity should not be adversarial. It goes both ways. I bet this was that one course nobody wants to pay for but have to because of an arbitrary "core" requirement.

RE: https://hcommons.social/users/christinkallama/statuses/115585211232840899
Show threadMatrew FileNov 21
@yume I believe this one goes with the side you choose to interpret. Of course setting an arbitrary trap in course evaluation to fail some already exhausted college students is of doubt for academic ethics, but submitting a slop (not necessarily produced by AI) for evaluation, that is clearly done in the most bodging way possible only to get a passing grade, from someone who clearly knows nothing, is a humiliation on the profession of the mentor. (1/3)
Show threadMatrew FileNov 21
This is just a thing of mutual respect: at least copy the response into a word processor and apply an overall Times New Roman, not just pasting screenshots from ChatGPT page by page - by all means, can't this much work be that hard? This is not a matter of AI; it's just a matter of personal qualities. One who chooses to rifle through a degree with minimal respect in higher education only to grab a generic office job and live a hopeless life deserves a disqualification.(2/3)
Show threadMatrew File
And no, this is not restricted to some annoying courses. I'm in my sophomore year, and some peers of mine managed to climb up with multiple LLM agents while they didn't even know how to install Python. I don't think that counts as an unnecessary requirement for an EE degree.(3/3)
Nov 21 at 11:30pmWeb
Show threadYumechi | ゆめち   Nov 21
@elfile4138 You missed my point (or you widened my argument into something I didn't say), I was talking about the history and literature courses that have met scalability bottlenecks on how to evaluate genuine progress while being ultimately utterly useless for an unrelated degree but forced into the curriculum.

It's a problem at least in the US. If the course is pointless and evaluation is mechanical, people take shortcuts.
Show threadMatrew FileNov 21
@yume The first sentence in the thread explained the divergence (sorry for the opinion detour, I just taken the liberty to express what I assumed was more important in the original post). As for the curriculum design, I'll admit that It's also present in Europe, and the low quality mentoring and evaluation are indeed a widespread problem even in critical degree related courses.
Show threadYumechi | ゆめち   Nov 22
@elfile4138 I personally prefer not to make sweeping judgements, my preferred way of argument is understand multiple things and multiple levels of argument can be true at once.

- the easiest argument to make is to endorse or discrete some other claims. I believe I don't have to prove or disprove any "benefit" of LLM to point out the hypocrisy and clickbait of this article. (Microblogging rarely exceeds this, a tweet cannot be a thesis, I didn't say who is right, I just say the article is not credible due to the speaker bias)

- then it is to make your own risk benefit analysis and make your own assertion (which is difficult especially when scope goes wide, I occasionally do them on my blog when I feel knowledgeable enough, this is not it)

- then it is to make a sweeping judgement ("AI should be banned for X"), which requires extraordinary evidence I bet nobody have.