A big part of understanding concepts (and being able to reason) is correlating with patterns you already know. Even though there are errors, LLMs are doing fair

A big part of understanding concepts (and being able to reason) is correlating with patterns you already know. Even though there are errors, LLMs are doing fairly well at pattern recall. Which makes me believe that we're perhaps one additive breakthrough away from another major leap in abilities.