For a long time, there have been two general approaches to AI. The first was focused on algorithms and heuristics, giving us everything from Eliza and somewhat-

For a long time, there have been two general approaches to AI. The first was focused on algorithms and heuristics, giving us everything from Eliza and somewhat-useful expert systems to Doug Lenat's Cyc project. Then came "AI winter" and everything (except some robotics) was pretty much on hold for a decade. Finally we got "cheap and deep" neural-net stuff as we know it today. I think we need both before we can reach true AGI. Whoever unifies them will have achieved the equivalent of grand unification in physics. And probably also created Skynet.