Poolside trains its Laguna coding models inside the same agent runtime it released for use with them. During training, agents attempt software tasks in isolated computing environments; their recorded actions and results are scored and fed back into model training. That exposes the models to sequences such as reading files, changing code and running tests. The released runtime therefore matters alongside the model: it supplies the tools and execution loop through which those learned capabilities become working software agents.
Terms in this piece · Glossary
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
tool use — A model's ability to call external functions — run code, search the web, edit files — instead of only generating text.
