This idea challenges the traditional notion that hardware and software should remain separate and stable. Instead, mortal computation envisions systems that evolve, degrade, and adapt in response to environmental stimuli, similar to living organisms. The significance of mortal computation lies in its potential to revolutionize artificial intelligence by embedding the principles of life—such as growth, decay, and self-replication—into computational models. This approach aims to create systems that are inherently adaptive, self-sustaining, and able to interact with their environment dynamically, much like biological systems. By adopting this framework, researchers like Geoffrey Hinton and others hope to develop neuromorphic, or brain-like, computing systems that blend analog and digital elements, leveraging the inherent "messiness" of hardware for more flexible and energy-efficient AI models(ar5iv)( RF SAFE® Radio Frequency Safe).
Historically, mortal computation has roots in biomimetic intelligence and cybernetics, but it gained prominence recently through discussions led by AI pioneer Geoffrey Hinton in 2022. He advocated for moving beyond the rigid, "immortal" nature of traditional digital computing, where software operates the same on any compatible hardware(Quantum Zeitgeist).
Key contributors include Geoffrey Hinton, who has been instrumental in promoting the idea of neuromorphic, analog-based systems, and Karl Friston, whose work on the free energy principle provides a theoretical basis for how these systems might self-organize and adapt in a way akin to biological entities(WhatsNewInPreprint).