Ihor Ivliev comments on An Analysis of Systemic Risk and Architectural Requirements for the Containment of Recursively Self-Improving AI