“If you think of neural networks as a software stack and not just a pretty good classifier, it becomes quickly apparent that they have a huge number of advantages and a lot of potential for transforming software in general.”
Reinterpreting the way you understand neural networks will completely change the way you think of software. And, according to Andrej Karpathy, the Director of AI at Tesla, it’s for the best.
Coined “Software 2.0”, which is written by neural network weights rather than humans, these weights will not replace Software 1.0 (much to the relief of today’s coders). Instead, Software 2.0 will help to lighten the load, taking on tedious tasks involved in creating the software behind visual and speech recognition, speech synthesis, machine translation, robotics, and games. Karpathy’s argument for the numerous benefits of porting complex programs into Software 2.0, despite its recognized limitations, should be enough to convince any coder to accept this new age transition.
Read more here.