Artificial general intelligence is back on track.

After the Google fiasco relating to misplaced definitions of sentient robots, it is heartening to see one of AI’s giants refocus us on real AGI. Yann LeCun is not resiling from his firm stance that deep learning is the sole answer to the development of artificial general intelligence. But he explicitly addresses Daniel Kahneman’s two modes of thinking, fast and slow, otherwise known as type 1 and type 2. His answer lies in the development of a deep learning architecture enabling a world model engine that can generate slow thinking. This is what we humans would experience as considered, reflective, effortful thinking. It gives us common sense, the ability to grasp causation as opposed to mere correlation and an understanding of counterfactuals. LeCun may or may not be right but this is a terrific step forward in the conceptual thinking needed to create AGI.

Link to article:

You might also like to browse other AI articles: