In the beginning people created the computer and coding. Now the code was dull and repetitive, boredom was running in the chip, and the spirit of people was hovering in front of the screen.
And people said, let there be AI, and there was fun. People saw that the AI was good, and they separated the connectionism from the symbolic AI. People found symbolic AI very auspicious and connectionism was set aside. And there was frustration, and there was no more funding – the AI winter.
Then people said, let there be some hidden neurons between the input and output layers. So people applied the backprop to neural nets and AI bloomed again. And people said, let ingenuity produce diversity: radial basis, convolutional, recurrent, and many other neural nets, according to their various architectures. And people saw that it was good.
But the data still needed pre-assigned labels, so people said, let there be unsupervised learning to separate the wheat from the chaff and automatically find patterns in the dataset. So, to complete the trinity of learning and allow agents to make decisions that maximize their cumulative reward, people also created reinforcement learning.
Then people said, let AI produce smart systems to their kinds: autonomous robots, the agents that move along the ground, air and water, each according to its kind; self-driving cars, vehicles capable of moving safely without a driver; and a bunch of smart home appliances: vacuum cleaners, lawn mowers, fridges and washers, each according to its kind. People saw all that they had made, and it was very good.
Thrilled with the greatness of their own creation, people said: let us make artificial general intelligence in our image, in our likeness, so that it may rule the world and all the creatures that live in it. And people waited, and waited, and waited… But what came was only: access denied, blue screen of death, device not ready, kernel panic, out of memory, and an endless stream of fatal error messages.
So, tired of waiting and annoyed by false AI prophets, people decided to have a few beers and talk about the latest projects they’ve added to GitHub. And it was so.