AI“Universal AI as Imitation” explores agency beyond utility maximization
The paper’s author suggests large language models point toward a different foundation for intelligence: learning patterns of behavior from the structure of interactions.
TLDR
Announcing “Universal AI as Imitation,” the author questions whether agency must be formalized as utility maximization. They suggest an alternative: intelligence as acquiring behavioral schemas—patterns of behavior—from interaction structure. The author describes the paper as investigating the limit case of large-language-model-style systems.
“Universal AI as Imitation” explores agency beyond utility maximization
The paper’s author suggests large language models point toward a different foundation for intelligence: learning patterns of behavior from the structure of interactions.
TLDR
Announcing “Universal AI as Imitation,” the author questions whether agency must be formalized as utility maximization. They suggest an alternative: intelligence as acquiring behavioral schemas—patterns of behavior—from interaction structure. The author describes the paper as investigating the limit case of large-language-model-style systems.
