hhmx.de

· Föderation EN Mi 07.05.2025 22:19:58

@nikatjef @cwebber how would it be possible to gauge student understanding with LLMs? Students can and have been copy and pasting the homework assignments to generate their work. There's no proof they are reading or understanding anything. If they are tasked with providing the prompts in order to prove their understanding, they will just provide fake LLM-generated prompts.

Also, I don't see how determining the students' prowess with LLMs has anything to do with the subject unless the subject is about prompt engineering.

Föderation EN Mi 07.05.2025 22:22:55

@vv
Yes, prompt engineering is critical for them as it will be their life within the next 10 years, so that is valuable in and of itself.

When a person looks into something they understand, their prompts will tend to be more focused on what they need, but also their prompts will generally show clear paths toward the information. Please note I am not talking about a single prompt, but the entire chain of prompts.
@cwebber

Föderation EN Mi 07.05.2025 22:23:28

@nikatjef @vv sorry in this house we don't do prompt engineering, only pronk engineering

Föderation EN Mi 07.05.2025 22:30:03

@cwebber @nikatjef @vv
"no prompt, no prank: only pronk!" (and proud)

Föderation EN Mi 07.05.2025 22:33:36

@cwebber @nikatjef @vv

Federated PRONK engineering. To be precise.

Föderation EN Mi 07.05.2025 22:25:12

@nikatjef @cwebber what a horrible worldview to think that the only thing people will be doing in ten years is trying to figure out how to best generate bullshit text