TrueNorth chips would probably be a competitive disadvantage. I found Yann LeCun's critique[0] compelling. In short, simulating the brain via spiking neurons

TrueNorth chips would probably be a competitive disadvantage. I found Yann LeCun's critique[0] compelling. In short, simulating the brain via spiking neurons is less efficient than just doing the floating point math like regular neural nets do. [0] https://www.facebook.com/yann.lecun/posts/10152184295832143 https://www.facebook.com/yann.lecun/posts/10152184295832143