Following is an offline email note that shows why nobody is using LLMs to check Social Media for notes that should be deleted or flagged as dangerous for one reason or another.  That is a very important application that would be worth a huge amount of money if it could be done accurately.

John

_____________

On Jun 15, 2024, John F Sowa wrote:

But Google AI does not understand satire.

On Jun 16, 2024 XXX wrote:

Or sarcasm, humor, anger, internet flame wars, stupidity, poetry, hunger or any other emotion or part of being intelligent or self-conscious.  Turns out that a gussied up autocorrect doesn’t handle subtlety.   Who knew?  No one knew it was that hard.

What is an emotion?