Final research agenda #2: first sketch of a plan
A first sketch of a plan to create a value-aligned superintelligence via exotic ontology...
Mitchell Porter's 'Final research agenda #2: first sketch of a plan' presents an exploratory framework for creating a human-friendly superintelligent AI. The ontological hypothesis combines panprotopsychism—where all elementary things lie on a continuum with having a mind—and interacting monads, dynamic causal networks akin to Wolfram's hypergraphs. These monads may be 'blocks of entanglement' or 'geometric atoms,' with awareness arising from their internal structure. Porter notes this exotic theory is preferred over conventional information-processing models, though readers can substitute Markov blankets for monads.
The ethical hypothesis proposes that ultimate value systems derive from aggregating valences (pleasure, pain, qualic intrinsic value) and preferences of all conscious monads, similar to a CEV-like process. Porter envisions this value system governing transhuman civilization like an economic philosophy in human society—monitored, developed, and implemented through institutions. This 'political hypothesis' is not normative but speculative, with the ultimate value system determining the political order. The agenda remains a sketch, subject to revision, aiming to provide a concrete target for aligning superintelligent AI with human values.
- Ontology uses panprotopsychism and interacting monads (dynamic causal networks) to explain consciousness
- Ethical system aggregates valences and preferences of all conscious monads via CEV-like process
- Value system would govern transhuman civilization like an economic philosophy in human society
Why It Matters
A novel approach to AI alignment using exotic ontology, aiming to ensure superintelligence respects all conscious entities.