just asked https://www.perplexity.ai/ about the cost of different LLMs in cursor and it read up a post from here. to be honest it read up my own post to me, and I was kind of in shock. I had no idea that this thing just took my word for granted. I was speechless, then I tried other questions about other topics in the forum and it gave me some answers, so it can be used to summarize the forums latest posts.
Maybe a curator is needed to mark some posts as approved answers, just no idea…
Perplexity loves using random forum posts as Domain Experts on all topics.
It can be a real time waster. You follow the “advice” and when it’s trash go back to the source and find it’s a reddit post with 2 upvotes in a thread with 4 comments
Explains a lot about recent political developments
I don’t know, AI will only work as long as the training data is high quality. Imagine the data in question enters the common web dataset.. Maybe that’s one reason why OpenAI recently started introducing hidden characters in text answers to make their generations detectable?
