Thanks for the shout-out @ruscion! I really like the idea of forming a cybernetic-intelligentsia of well-structured arguments about the path forward for AI. It reminds me of the early days of /r/Futurology.

One of the reasons I'm "optimistic" for an AI-heavy future is that right now the top talent in deep learning is getting paid like NFL quarterbacks.
http://www.economist.com/news/business/21695908-silicon-valley-fights-talent-universities-struggle-hold-their

Respectively, based on the advancements in three-dimensional integrated circuits, Quantum Computing (both Annealing and Universal models), and in combination with Artificial General Intelligence, I would argue we are rapidly approaching a new era.

Without a doubt. We're already entering a new era or sorts, the Chinese supercomputer Sunway TaihuLight is pushing 125 PETAFLOPS. This is the lower end estimation of the human brain neural connections. Things like the Human Connectome and the Allen Brain Institute are steam-rolling ahead in ways to model the brain.

Ultimately - I think we get AGI in one of two ways:

• 1. General laws of cognition - something approximate to the way airplanes can fly, but as a general mathematical law about self-organizing neural networks.
• 2. Direct brain emulation - we model the electrolyte and neurotransmitter channels of the entire brain completely.

I think we'll get both types. In my series 'The Star Steaders' I'm describing this difference as "Wet Cogs" versus "Dry Cogs" where Wet means a simulated brain.

More on that soon! And yeah, even quantum computing will probably help too. I'm not super impressed by the adiabatic stuff that D-wave is doing, but if they can figure out how to do Topological QC / braid groups - all bets are off!

Keep up the awesome posts! Maybe we can find a good shared tag to get started. I wish they gave us more than five!