[tmvphil]

LLM's are reducible to matrix multiplication in the same way that human cognition is reducible to physics. Reducibility does not imply understanding.