▲ | crazygringo 3 days ago | |||||||
Honestly, this is the type of stuff ChatGPT is really good at. Explaining overviews of a field, avoiding math if you want, focusing on concepts and explaining different schools of thought. As long as you're sticking to the well-established stuff, it tends to be quite factually accurate. I think it's really underrated as a resource for good high-level overviews of fields where those overviews otherwise don't exist at all, are overly technical, or the existing overviews have a lot of author bias. | ||||||||
▲ | jzl 3 days ago | parent [-] | |||||||
Ugh, I couldn’t disagree more. Sure, LLMs can generate some really nice introductory summaries of topics. But, so far at least, they can’t even hold a candle to brilliantly written books and long form articles. Consider the classic book Cosmos. There is more insight into the universe in any few pages of that book than could be gathered by reading even a thousand ChatGPT results. | ||||||||
|