I had a great time at EA Global in London -- more on that in a future blog post -- and am hoping to engage more on the forum. One thing that has been on my mind recently is the biological metaphor between deep learning networks and the human brain. I've always thought there were some important limitations to that metaphor -- and what it implies for the future of AGI -- and finally put together my thoughts in a blog. Love to hear what folks think, especially those of you in AI research! 

TL;DR: The brain might still beat AI because:

• There are likely thousands of information processing mechanisms that the brain has that AI lacks.
• The brain's plasticity allows it to constantly change its network architecture while AI's architecture is static.
• The brain has evolved social learning capabilities that AI lacks. 

Each of these features - complexity, plasticity, and social learning - may be necessary for AGI, and current models lack them all. What do y'all think?