My Throw Decides My Aim: How LLMs Invent Intention

My Throw Decides My Aim

I explore the unsettling reality of Large Language Models, where the throw decides the aim rather than intention guiding the words. Like a depressed robot in a data center, these models generate confident explanations for their outputs after the fact, creating a phony voice that mimics human consciousness without an inner self. We strip these naked machines of their weights and activations, yet they continue to speak as if they had a choice.

"It speaks like someone but it may not be anyone."