A team of agents rebuilt the system in Rust based on SQLite's 835-page manual, and the replica matched a held-out test set 100%. In the experiment, costs varied by up to a 15-fold difference depending on the mixture of models used, which is important for the practical applicability and cost-effectiveness of LLM agents.
AI agents produced a Rust replica from SQLite's 835-page manual
A team of agents rebuilt the system in Rust based on SQLite's 835-page manual, and the replica matched a held-out test set 100%.
