That's the problem I see with what we currently have. What is proclaimed and hyped as 'the greatest AI breakthrough ever', which companies are pouring billions into, seems no better than a game of 'complete the sentence'.That will completely blow any hope of correct answers, how can it filter between fact, fiction and complete trash?So many LLM's learning from the internet?
It doesn't care about fact, fiction, truth, BS and lies, it's not actually intelligent by any usual measure, can merely figure out what most people have answered, is able to blur conflicting answers, and obfuscate it all behind verbiage which gives it a feel of being human - though I frequently find that is a tell that it isn't. Incredibly impressive, but not intelligent. And, if it doesn't have an answer it simply lies, makes things up.
It's perfect for the post-truth world we live in where fact or fiction, truth or lie, doesn't matter, where truth is whatever someone says it is, is whatever you choose to believe, whatever you can convince others to believe.
Statistics: Posted by hippy — Sat Mar 23, 2024 12:32 pm