This is from a talk given at Defrag 2014. One of the (few) advantages of being in technology for a long time is that you get to see multiple tech cycles beginning to end. You get to see how breakthroughs actually propagate. If all you have seen is a part of the curve, it’s hard to extrapolate correctly. You either overshoot the short-term progress or undershoot the long. What’s surprising is not h