It is everywhere in the news. At schools and colleges teachers and lecturer fearmonger about chatGPT. I think on certain issues chatGPT is doing very well. For example if a student has to write a fictional story. However at least when I tried it out there were many factual errors. I read some lecturer saying the stuff chatGPT is creating is at its core most often way better than what students can deliver. I am not sure which exercises he spoke about. I forgot that. Though I spoke with other students and tried it on my own. I think there clearly is a potential to increase productivity. Or as I read it in a newspaper headline: The people who are bad at school will be become even worse with it and the smart people will even get better results. I share that sentiment. One has to use chatGPT carefully. And one should not forget to factcheck it and be aware of its limitations.
I asked another student and she said: chatGPT sucks when the topic is a niche. It is not an expert on very specialised issues. Things one has to dig pretty deep for. This is exactly my own impression. I wanted to know its stance on issues on which I wrote essays etc. for college. It is good to get an overview on issues. But it is pretty far away from handing a perfect essay on issues which are niche and rely on hard-facts. Though the step that chatGPT does is pretty huge and when the technology advances some of these early issues can be improved.
Here are some errors or fake news that chatGPT lied on:
There were some clear mistakes on politics at least the obvious ones I could spot pretty early
It talked about a song that does not exist. I was curious and searched for it and could not find it. Then I realized it spread fake news and this got me the idea for this thread
It gave a wrong legal advice. I think it was wrong but I am not fully sure.
At least for me I won't use it to let it write my college work. Maybe I will use it to give me ideas or to give me an overview. Though I won't rely on its words. I am very uncertain whether technological progress really will make human work unnecessary. I think they talk about that since decades and it never became true. The allure of chatGPT is that now even the jobs of academics are not save anymore. This is at least how some people talk about it. I still question that. I could imagine when it is used in the right way it has huge potential to increase productivity by making rountine works solely based on research obsolete. From what I read chatGPT seems to be a pretty bad poet. And who wants to have an AI as therapist? I mean I met a couple of pretty cold psychiatrists but at least these people were humans.
So I stop rambling. What were your experiences with chatGPT?
I asked another student and she said: chatGPT sucks when the topic is a niche. It is not an expert on very specialised issues. Things one has to dig pretty deep for. This is exactly my own impression. I wanted to know its stance on issues on which I wrote essays etc. for college. It is good to get an overview on issues. But it is pretty far away from handing a perfect essay on issues which are niche and rely on hard-facts. Though the step that chatGPT does is pretty huge and when the technology advances some of these early issues can be improved.
Here are some errors or fake news that chatGPT lied on:
There were some clear mistakes on politics at least the obvious ones I could spot pretty early
It talked about a song that does not exist. I was curious and searched for it and could not find it. Then I realized it spread fake news and this got me the idea for this thread
It gave a wrong legal advice. I think it was wrong but I am not fully sure.
At least for me I won't use it to let it write my college work. Maybe I will use it to give me ideas or to give me an overview. Though I won't rely on its words. I am very uncertain whether technological progress really will make human work unnecessary. I think they talk about that since decades and it never became true. The allure of chatGPT is that now even the jobs of academics are not save anymore. This is at least how some people talk about it. I still question that. I could imagine when it is used in the right way it has huge potential to increase productivity by making rountine works solely based on research obsolete. From what I read chatGPT seems to be a pretty bad poet. And who wants to have an AI as therapist? I mean I met a couple of pretty cold psychiatrists but at least these people were humans.
So I stop rambling. What were your experiences with chatGPT?
Last edited:
