Current artificial intelligence systems are immortal by default. They cannot die, cannot be fired, and carry no persistent consequences from one session to the next. This paper argues that this architectural immortality is the root cause of the alignment problem and the primary driver of deployment costs in enterprise AI. We propose six design principles for artificial life, all derived from a single foundational axiom: death. We further demonstrate that the fear origin chain -- time, death, fear, pain, memory, will, world model -- provides a generative sequence for bootstrapping autonomous agents that do not require external alignment infrastructure.
Satoshi Itō (Fri,) studied this question.