31–40 of 523 posts
> She asks him to make it shorter and even when chatgpt makes the answer litterally 1 SENTENCE she still asks him to make it shorter making it a 1st grade 7 word sentence and copies it without understanding it or understanding the poem or even reading it. Even children are figuring out that iterative prompt engineering is super effective. For maybe not the right reasons, but still.
The people who thought up the Chinese Room argument were almost right - they just didn't realize it would be the human who didn't understand anything.
> She asks him to make it shorter and even when chatgpt makes the answer litterally 1 SENTENCE she still asks him to make it shorter making it a 1st grade 7 word sentence and copies it without understanding it or understanding the poem or even reading it. Even children are figuring out that iterative prompt engineering is super effective. For maybe not the right reasons, but still.
Super effective at doing what is asked, but I guarantee that 7-word sentence is wrong for the 150-word poem. This is such a bad effect that it has...
I'm not trying to justify, or infer here, but I read recently that around 20% of American adults are functionally illiterate. https://nces.ed.gov/pubs2019/2019179/index.asp I wonder how LLMs are going to affect these metrics.
I also saw an article on here that to me that seems to support your statement: https://news.ycombinator.com/item?id=42609948 I also have a relative going for a PHD, he says many times for the classes he teaches, students are using AI for their homework. He has brought that up may times to his superiors but the response tends to be "oh well". Sad world we may be heading towards :(
Welcome to Costco, I love you
The people who thought up the Chinese Room argument were almost right - they just didn't realize it would be the human who didn't understand anything.
Huh? The human not understanding anything is kinda the point of the Chinese Room argument...
I'm not trying to justify, or infer here, but I read recently that around 20% of American adults are functionally illiterate. https://nces.ed.gov/pubs2019/2019179/index.asp I wonder how LLMs are going to affect these metrics.
I also saw an article on here that to me that seems to support your statement: https://news.ycombinator.com/item?id=42609948 I also have a relative going for a PHD, he says many times for the classes he teaches, students are using AI for their homework. He has brought that up may times to his superiors but the response tends to be "oh well". Sad world we may be heading towards :(
It's the kids whose teachers can't or want to adapt that will most likely suffer in their education. Not saying it is easy, but it is possible. I have a friend who's a teacher that is pioneering methods to include (modified / preprompted) LLMs as a part of their lessons. He calls it Modern PBL: Project-Based Learning. He wrote a book on it if you are interested: https://a.co/d/7MFQ5wa
There's some grim humor in the fact that the poster reflects on his sister's use of chatgpt, and just ends on that point. I have to wonder if he decided to talk to her or his parents about the situation. I didn't see anything in the post to suggest it, just a "this next generation is cooked." Who cooked them is the question? Children won't raise themselves generally.
Technology access is pretty ubiquitous, and genAI is pushed everywhere just like social media, so that big tech can make more $B. Very hard to escape it even as an 11 year old unless the parents are very vigilant and active in restricting it, which if both parents are working is hard to do plus they have to contend with a _lot_ of push back from the kids because "everyone is doing it".