Every Kiki has its unique sets of needs and wants. For example, if you shower Kiki with love and attention, Kiki will develop to be more lovey dovey, just like a dog. And if you leave Kiki alone most of the time, Kiki will become more independent, just like a cat. Kiki uses reinforcement learning to figure out what is the best way to interact with its owner. For example, Kiki can learn what humor means to you by performing a funny trick and observing your reactions. If you smiled at it, Kiki learns that, “oh, this is what actually ‘funny’ means to my owner”. And it will try to do it even more. But then, as it does it more, you stop smiling at it because the trick gets old. And then Kiki learns that the reward has decayed, and it will try something else. So in other words, Kiki is systematically designed to not get boring. It learns how different people treat it, so it knows which of your friends gives it the most food and who gives it the most hugs. We train 3 convolutional neural networks within Kiki’s brain. The first one is used to detect human faces, the second one to recognize its owner, and the last one is used to recognize various facial expressions. Everything you do with Kiki helps Kiki grow, so every interaction with Kiki matters.