Part 11/17:
Bach advocates that self-organizing architectures—like neural cellular automata or wave function collapse algorithms—might underpin artificial consciousness. Instead of static neural networks, he envisions systems capable of developmental learning, dynamically forming hierarchical scripts, context-based hypotheses, and perception models, similar to biological evolution.
He explores hierarchical request-confirmation networks, where units engage in dynamic scripting and cooperation, allowing perception and cognition to operate via hypothesis testing. This bottom-up/top-down approach enables systems to interpret complex scenes, resolve ambiguities, and learn from interactions, gradually developing awareness.