Every interface imposes cognitive load. The question isn't whether to add load, but whether what we add serves human goals. Most knowledge tools fail this test—they require users to learn complex organizational schemes, navigate nested hierarchies, and maintain mental maps of where information lives.
How LLMs Model Mental States
Large language models exhibit surprising capabilities in reasoning about beliefs, intentions, and knowledge states. We explore what this means for building systems that truly understand context.
Read article→
