Live data from Hacker News

My little sister's use of ChatGPT for homework is heartbreaking

old.reddit.com

521–523 of 523 posts

Re: My little sister's use of ChatGPT for homework is heartbreaking

#521

Earlier quoted context omitted.

Why's that?

The student is growing up in a world of technology. How can the homework be designed leveraging LLMs? Can some curriculums be accelerated assuming LLMs provide a new base? What is the goal of teaching? Memorizing formulas or learning to think? Do students gain any value writing a book report or is there more value writing an essay about how a book relates to a students personal story? Anything that is formulaic in na…

I think there’s something to be said about building up a personal knowledge base and automaticity with retrieving basic information to free up your working memory for more complex problems.

I’m still grappling with what this means now that LLMs have large enough context windows to ingest books.

Re: My little sister's use of ChatGPT for homework is heartbreaking

#522

Earlier quoted context omitted.

How does that work with disability accommodations for students that have to be at home?

A strict system does not contradict making accomodations for those that need them. All countries do it differently but in several European countries you even have graduation certificates that may indicate if you followed a special curriculum or had some accomodations, if relevant for the final result/report.

I meant: how is actual testing carried out?

Re: My little sister's use of ChatGPT for homework is heartbreaking

#523

Earlier quoted context omitted.

Do you train a good copy typist by giving them a xerox copy machine? Can you even be a good copy typist if the only documents you've ever copied is from a xerox machine? No and nobody cares because copy typists don't exist any more.

But xerox machines are good at their jobs compared to LLMs

LLMs are 2 years old. Xerox machines took much longer than that to become reliable and economical. Are you supposing that AI systems have peaked and will no longer improve?

I can find threads from 12 months ago claiming LLMs will not be able to do things they now do all the time.

Post reply on HN