Why reinforcement learning?

In 2014, Google acquired a London-based startup named DeepMind for a whopping $500 million. In the news, we read that they had created an AI agent to beat any Atari game, but the main reason why Google paid so much to acquire it was because this breakthrough was a step closer toward general artificial intelligence. General artificial intelligence is referred to as an AI agent. It is capable of doing a variety of tasks and generalizing just like a human. When it surpasses that, that point of singularity is termed, artificial super intelligence. At present, the work done by the AI community is what we term, artificial narrow intelligence, where an AI agent is capable of acing a couple of tasks but not able to generalize over a variety of tasks. 

DeepMind published their paper, Human Level Control through Deep Reinforcement Learning in the...