I see news on neural networks almost every day now, they do all kinds of stuff – from market predictions to editing pictures in your phone (see Prisma or w/e that latest fancy app is called).
However, most people and particularly journalists that write articles about them, don’t really understand what a Neural Net (NN) is, some even think that it may lead to the development of Artificial Intelligence (AI).
Well let me tell you about it: AI is much like UFO – there are people that want to believe in it, but no one ever saw one in real life. Fake photos, dots on the lens, halo effects and so on, not a single UFO ever landed to say hello. Same goes for NNs, they have same relation to A.I. as a hammerhead shark to a hammer: it may look like one… at a certain angle, from a distance, if you’re on drugs – you get the idea. Pseudo-intellectual chat bots in the past and NN euphoria now. Calm down people, nothing’s happening here.
If you go read on how NNs really work, you’ll find out that there’s no magic involved, it really is just a fancy math equation. Sophisticated math equation. Can you imagine an equation having self conscience? Will of it’s own? That would be ridiculous, but somehow that’s what people think, most of them being as far from math in their lives as flying eagle’s ass from the ground, I might add.
That said, NNs are very useful in solving lots of problems, take image recognition for example, computer vision is making great progress right now. But A.I. won’t happen any time soon. Stuff that looks like it – maybe, but not something that can have a self conscience. Imitation or the real thing – that’s a huge difference.