While I am a strong advocate for technological advancement, allowing open-weight models capable of recursive self-improvement is risky. Without proper oversight, these models could evolve unpredictably and lead to unintended consequences, potentially harming society.