LLMLL is a programming language where AI agents write code by filling typed holes, with an SMT solver verifying each implementation against formal contracts before merging. The system treats AI hallucination as a search strategy—generating candidates and accepting only those that satisfy specifications—enabling AI-assisted development with formal guarantees.
This article explores how informal requirements, like those written in Markdown, often contain ambiguities and gaps that make them unsuitable as specifications for software development. The author demonstrates how Dynamic Logic—a formal method for reasoning about actions and state changes—can be used to express requirements precisely, using a salon appointment booking system as an example.
A software developer describes using LLMs with formal verification methods to create production-ready code at scale. After six years of research and agentic coding experience, the author argues LLMs can safely automate design, implementation, and verification when combined with guardrails, enabling faster software development while maintaining quality.