Superhuman Articulacy as an LLM Safety Target
TL;DR: Current LLMs are bad communicators relative to their agentic capabilities. I claim that articulacy is useful (and perhaps necessary) for AI safety and suggest a path for improving articulacy.Briefly: a theory for articulacyFrequently, LLM agents miscommunicate with their human operators, such as when they write documentation or respond to queries about their activity during a coding session. Any given communication failure can be ascribed to either or both of these two factors:ArticulacyI
By Dylan Bowman