That’s exactly it. Just ask a child from the West and from Asia (without knowing the other’s culture) to draw a dragon. One will make a fire-breathing lizard with wings, and the other a water-breathing snake with legs and whiskers. Why? Because that is what these people have as a frame of reference. There is also a reason why, for example, all mythical creatures are composed of properties of real beings, it has been said that only properties exist in real beings, and they are said to be a combination of real beings.

It says above: “Yes, they don’t need 200,000 new images to figure out the new arm position.” But we did. However, we have been seeing those “images” through our eyes all our lives.

The neural network (or actually artificial neural network, ANN) that AI often runs on (for example, LLM is a type of ANN) ultimately depends on how we think our brains work. We currently have more computing power than a CPU, which means we can’t be matched yet. But even our emotions are made up of chemical reactions (hormones) and electrical impulses in our nervous system (brain), which in theory can be computed with enough computing power. It is only the precise combination of these things, as well as our experiences, memories, and the connections that form in our brains, that varies, which is what makes us all different. I really wonder if we (the current people who are now talking about AI) will ever see the time when this actually happens, because we are still very complex and still a long way from calculating it all. But our humanity is not half as irreplaceable and unpredictable as we would like to lead ourselves to believe (nor as much superior to other animals as we would like to lead ourselves to believe).

This potentially makes AI “scary” to us humans. But this is exactly what makes AI great.

This, aside from the fact that AI is often an empty buzzword these days, and is basically “anything perceived as intelligent” when it comes to software. It is much more interesting to talk about the technologies used than the term “artificial intelligence”. Much of today’s artificial intelligence is nothing new at all. Even the mathematics behind artificial neural networks dates back to the late 18th century, and computer networks have been around as early as the 20th century. The only difference that suddenly makes people think a lot about it is that, first of all, it is now popular and therefore widely marketed. Secondly, it is now starting to become powerful enough to have actual human-like outputs that make one sense the approaching “danger”. Especially since we have always been the best “calculators” and thus have reached the top of many food chains

[Reactie gewijzigd door Cambionn op 2 januari 2024 15:29]