For the past few days, I've been testing an impressive open-weight AI coding model called Flint. It suggests that fully recursive self-improvement, or RSI, is getting closer to reality.

Flint is the first offering from Recursive, a company that aims to use RSI to advance AI and other areas of science using AI. The CEO and co-founder, Richard Socher, says that nearly every aspect of the model pipeline was actually done and improved upon by AI.

The Eureka Machine, Recursive's recursive AI system, automated most elements of the model's development, enabling a small team to build a world-class model in a few months. Flint outperforms other models of similar size and architecture on coding tasks, according to Recursive.

AI is already used to help build, test, and improve other AI models. This can speed up the process of refining a model and testing new ideas without the need for as much human input. It is possible to create a loop through which AI decides what improvements to make and then begins another round of improvement.

Recursive sees these fears as more sci-fi than reality and believes legislation to prevent RSI is an unnecessary impediment to progress. Flint was built using a system incapable of escaping human control, and the Eureka Machine itself could achieve full recursive self-improvement, but this would still be overseen by humans who could stop the process if it goes awry.