“Sakana AI's research philosophy embraces Ken Stanley's concept of "open-endedness" to explore ideas that may be under-resourced in the broader ML community.”