DeepMind scientists: Reinforcement learning is enough for general AI (bdtechtalks.com)

edit the title -- "AI" not "A" at the end

Cut the middleman:

https://www.sciencedirect.com/science/article/pii/S000437022...