Friday, June 13, 2025

Go figure

New Apple study challenges whether AI models truly “reason” through problems

Librarians have been saying this for 25 years. 
They are fancy sentence generators.   They operate on statistics.  They calculate that their answers are correct based on frequency of examples in the training set.   They are  a fancy autocomplete program.  
They do not understand cause and effect.   They get it confused all the time.
They only find patterns.
They cannot be creative.