
Rethinking LLM Programming
A Theoretical Framework Based on Automata Theory
The Ann Arbor Architecture introduces a paradigm shift in how we program Large Language Models by reconceptualizing them as automata that process unified natural and formal languages.
- Treats LLMs as automata that should be programmed in the languages they naturally accept
- Challenges traditional software engineering's separation of programming and natural languages
- Introduces Postline - a platform implementing this new architecture for agent-oriented programming
- Bridges theoretical computer science with practical LLM applications
This research matters because it provides a formal foundation for prompt engineering, enabling more systematic and effective development of LLM-powered applications across industries.