alt Hacker News> grading is a menial task that is below the station of a professor
Isn't there a P != NP element to this? Doing the work is hard. Checking it is not.
If a student honestly attempted a problem and failed, diagnosing the failure may be worth the professor's time, but given the bunchiness of undergrad (and frankly, any junior learning) mistake varieties, chances are the TA can identify the issue as well.
Replies
There is, but AI really turned the tables here. With human students you can gauge their "priors" and grade it fairly easily and like you say the mistakes are bunchy. With AI the priors are so different and their steering and everything ends up in a deep dark forest where no one knows whats going on. And it becomes a whole first-principles effort to grade it. Imagine if everyone in class was that weird genius who solves things in a roundabout way that you end up appreciating 2 days later. In person exams have become a must almost.
Sort of, but not really. For multiple guess type tests, there is that element. However, other types of tests and particular homework, there's not that as much. We're not interested so much in did you get the right answer as did you follow the right techniques.
At least this is about math where there is usually a single right answer. For things like English where the question is, did you get the right points? There's should be debate over even what is the right points in any large work. And so if you're convincing in your argument of what is the important points, you can pass homework.
Edit, fix autocorrect