New ICML paper with @SchmidhuberAI on how we can think of Interestingness, how it connects to open-ended learning and recursive self-improvement, and how algorithmic information theory provides concrete insights! https://inductive-interestingness-2026.github.io/future-compression-progress?v=2