[[Tony Seale]] post on LI articulating the [[Semantic agent]] pattern. > On one side: powerful model inside powerful harness > On the other: powerful data inside powerful ontology > > Put them together and you get the Semantic Agent. It doesn't just generate. It starts to understand. ([[LLM]] + Harness) + ([[Ontology]] + Data). That's the whole pattern. Everything else is scaffolding - useful for a while, but scaffolding comes down. ## Comments > An agent that inherits your organizational confusion **at scale** is just a faster way to be wrong. ^agents-inherit-confusion