We start today’s installment with the very cliffhanger sentence yesterday’s installment ended with: Rodney saying “Yeah, let’s talk about deep learning.” We proceed to do just that. For anyone giddy about the glittering newness of neural networks and the deep learning systems they power, Rodney points out that this work began in 1943.
This leads to an argument similar to yesterday’s point about self-driving cars regarding the importance of knowing a technology’s full history before handicapping its future. Rodney’s basic point is that deep learning is an overnight success that required 70 years to percolate. So the next giant breakthrough could be further off than we think.