They can do fun and interesting stuff, but we keep hearing how they’re going to replace human workers, and too many people in positions of power not only believe they are capable of this, but are taking steps to replace people with LLMs.
But while they are fun to play with, anything that requires a real answer, but can’t be directly and immediately checked, like customer support, scientific research, teaching, legal advice, identifying humans, correctly summarizing text - LLMs are very bad at these things, make up answers, mix contexts inappropriately, and more.
I’m not sure how you can have played with LLMs so much and missed this. I hope you don’t trust what they say about recipes or how to handle legal problems or how to clean things or how to treat disease or any fact-checking whatsoever.
>I’m not sure how you can have played with LLMs so much and missed this. I hope you don’t trust what they say about recipes or how to handle legal problems or how to clean things or how to treat disease or any fact-checking whatsoever.
This is like a GPT3.5 level criticism. o1-pro is probably better at pure fact retrieval than most PhDs in any given field. I challenge you to try it.
But while they are fun to play with, anything that requires a real answer, but can’t be directly and immediately checked, like customer support, scientific research, teaching, legal advice, identifying humans, correctly summarizing text - LLMs are very bad at these things, make up answers, mix contexts inappropriately, and more.
I’m not sure how you can have played with LLMs so much and missed this. I hope you don’t trust what they say about recipes or how to handle legal problems or how to clean things or how to treat disease or any fact-checking whatsoever.