Observation 1: If you don’t want AI Responses, don’t reward it

The really frustrating markts that I’ve lost in courses this semester have come from times when I wrote reading responses myself and did not use AI.

Seems like some TAs are expecting responses that are as polished as AI responses.

Well, be careful about what we ask of others, because we might just get it…

Observation 2: Course staff seems clueless about how powerful AI is, partly because we are all hiding that from them.

I see people completing what’s expected to take 3-4 hours in a single prompt. Course staff seems ignorant about this. (Or, even worse, they don’t care as long as they are getting paid.)

Collectively, we (students) are pretending to be much less AI literate than we are, so people don’t assign more work to us.

Case In Point:

I made more progress with Claude on one of my group final projects (AI Encouraged) in ~2 hour (as I was multitasking on other psets) than a team of 5 made in 3 weeks. We were all pretending to do work in order to get the marks from the weekly project check-ins. If we did not have the psets, we’d have finished the project by now.

Observation 3: Some papers are worth reading, and AI summaries just do not help

However, some papers are not worth reading. Reading the AI summary is a good way to convince myself that I’ve read it, so that I can go on with reading the papers worth reading.

Observation 4: Documentations of a lot of AI products are impossible to read

I find man pages hard to read because they are so dense. These AI documentations are impossible to read because they have no information.

Please, just give us the prompt to generate our documentation if you’re not bothered to write a good one yourself. No shame.