AI Safety, Uncontrollability, and Existential Risk: A Critical Review of the Superintelligence Problem | Synapse