Ripper piece.

Agreed. I didn't finish my first PhD attempt, and that was in no small part due to the fact that I had to wrestle with being expected to treat it as a full time job, while not earning enough money to pay my rent and bills. In the end, all the work I had to do gave me more options than the PhD did.

I think it works better with programmatic approaches, because the important assessments should be pushed further up the Blooms Tax and have more 'correct answers' as a result. A greater variety of versions of success = a greater variety of paths to that success.

I'll be working on some stuff to differentiate an AI agent from a human student - I already readily differentiate students from human agents (contract cheating providers). Watch this space.

Talk to @ccguerilla.bsky.social about it. He recently did one for #AcademicIntegrity

Ding, ding, ding. No need to carry out assurance of learning with the test. Suddenly cost effective again.