A popular way to explain how current LLMs work is to say that “all” they do is predict the next most likely word in a sentence. From one perspective, this is correct. Trained on all human language, the LLMs distilled … Continue reading →
Your “continue reading” link doesn’t work. Anyway, LLMs aren’t “smart” the way the word is used for smart humans. As software they’re smart in the sense of being advanced, complex software. LLMs use sophisticated math that weighs millions or billions of factors to compare prompts with massive databases of human-created content and determine the most likely response a human would make. They don’t possess the kind of intelligence that created all that human content, they just simulate it using a completely different process.
Your “continue reading” link doesn’t work. Anyway, LLMs aren’t “smart” the way the word is used for smart humans. As software they’re smart in the sense of being advanced, complex software. LLMs use sophisticated math that weighs millions or billions of factors to compare prompts with massive databases of human-created content and determine the most likely response a human would make. They don’t possess the kind of intelligence that created all that human content, they just simulate it using a completely different process.