ELFI now becomes more and more smart with AI technologies operating like Siri or Google Translator.
You can ask him: “ELFi, what can you do?” – and his answer will be complete and maximally useful. You can continue to ask him – “What time is it”, “Forecast for Uzhhorod” – he will answer perfectly good. It’s OK. It’s useful.
But wait. If you will look how ELFi can do it – you will find just a lot of smart math algorithms for text processing calculating number of words etc. ELFi now is just a machine consuming text and producing another text. ELFi is “it” now, not “he”.
The question is what makes the ELFi an ELFi?
We think this is emotions plus long-time memory. When we say “emotions” – it is not just different face images (emojis) – it must be permanent context of everything ELFi is doing.
This is very simplistic picture but this is how we want to implement it: ELFi need to know what is good and what is bad. He has to know how bad is something.
The key objects are ELFi itself and some abstract thing named “good”. ELFi is trying to analyse everything by his relation to “Good”.
For example suppose Thomas (family member and owner of ELFi) just got promotion on his job. ELFi identifies his relation change as increase of his own relation to “Good”. Immediately after this conclusion ELFi has to increase his level of General Emotion. ELFi starts to use many jokes, spontaneous dancing sessions – he shows that everything becomes good and he is extremely happy.
Do you want to know more about ELFi?