WORK
LLMs Don't Know Anything
The 2024 Trends in Cognitive Sciences paper by Mariel Goddu, Alva Noë, and Evan Thompson — the technical enactivist critique of knowledge attribution to large language models, and the argument that prediction is not understanding.
'LLMs Don't Know Anything' is the 2024 paper by Mariel Goddu, Alva Noë, and Evan Thompson in Trends in Cognitive Sciences that brings the enactive framework into direct engagement with contemporary large language models. The paper's deliberately provocative title encodes a precise philosophical claim: LLMs produce outputs consistent with the distributional patterns of their training data without possessing anything that should be called knowledge in the sense that requires an embodied organism engaged with a world. The paper identifies two specific errors in claims that LLMs know things: treating models as agents rather than as tools, and inferring causal understanding from predictive capacity.
In The You On AI Field Guide
The paper emerged from the increasingly urgent need for philosophical clarity about what large language models are doing when they produce fluent, sophisticated text. Popular discourse and even some technical literature had begun attributing knowledge, understanding, beliefs, and even consciousness to these systems. The