Steve Taylor

This paper introduces Daneel, a scalable architecture for building consciousness into Artificial Intelligence (AI) systems allowing for Artificial Superintelligence (ASI) devices or machines, based on the Executive Continuous Thread (ECTC) Theory of Consciousness. ECTC expands upon Global Workspace (GW) theory, which emerged in Cognitive Science in the early 2000s and posits a "blackboard" system for distributed knowledge processing. While GW theory has effectively predicted conscious behavior in biological systems, Daneel offers not only a clear definition of consciousness, but the first practical, buildable framework for implementing synthetic sentient AGIs. It employs a cell-based Hierarchical Manifold system that addresses challenges in intelligent agent management, learning, planning, knowledge storage and transfer, modeling the real-world, and real-time decision-making. Daneel incorporates feedback paths for learning through enhanced neural network structures and scalable deterministic machine learning engines with parallelism, allowing for adaptive behaviors, real-time reasoning, and grounded intelligence. This framework shares similarities with biological models, such as Neural Darwinism and Connectionist perspectives on brain development. Brain imaging shows that consciousness is linked to widespread cortical activity and continuous integration of sensory information, while unconscious states involve only localized activity with reduced overall brain and sensory processing. These findings support the principles underlying Daneel, offering a promising foundation for developing conscious intelligent machines. At the conclusion of this paper, we will offer a definition for intelligence and consciousness that definitively supports the design for building Daneel.

A pioneering leader in AI and Synthetic Sentience, Steve is known for his research in Artificial General Intelligence (AGI) and Computer Design. His career spans C-level roles where he led major product developments in parallel processing, computer design, and AI. He currently is Chief Scientist at Deriv8agi, Inc. (deriv8.ai). He brings deep expertise in Systems, Intellectual Property, AI, ML algorithms, architectures, and CPU design. He has solved the Master Algorithm - a 40 -year problem in machine learning that has alluded some of the best minds in AI. He is an author of multiple books on robotics and AI.