Goertzel’s own bet replaces the neural net as the core data structure with a self-modifying knowledge graph that fuses neural, logical, and evolutionary learning.
The basic data structure isn’t a neural net, it’s a huge dynamic self-modifying knowledge graph living in RAM across multiple machines. You can put neural nets in there, you can put logical propositions in there and do theorem proving, and you can put program code in there and evolve it with genetic programming that simulates evolution by natural selection. Then you wrap a goal-driven cognitive architecture around it.
Source: Rethinking the AGI Race, with Benjamin Goertzel (eCornell Keynote)