[eonmist]

It is an amazing process being shown here. It shows how technology, especially AI, opens new ways of learning. Examining the process, it makes sense that he was able to create it. This showcases interplay between modern technology and pedagogy, while also focusing on an explorative learning process, which, in modern pedagogy, is highly valued.

It seems clear the understanding of the game idea stems from experience with Pokemon, which introduces children to numbers and the concept of hits, damage, health. Thus it is natural he would have such a game idea as an entry point.

Abstract thinking usually develops around the age of 12. My question was how he was able to make symbolic representations of abstract entities (functions) at that age. I think this is where AI shines, as it bridges the gaps in understanding, thus lifting the burden of solving those gaps instantly and instead allowing for the fun explorative process of engaging with the problem at the current level of understanding -- keeping you at the "almost there" feeling, thus continuing is engaging if not irresistible.

In pedagogy, the zone of proximal development "is a concept in educational psychology that represents the space between what a learner is capable of doing unsupported and what the learner cannot do even with support." (Wikipedia) 20 years ago, this would probably be too difficult even with support from a parent. The parent would have had to do much more of the heavy lifting to allow the child to not be too distracted by syntax and implicit conceptual relationships. For example, in setTimeout(function..., it is not intuitive, even for many adults, to declare a function within a call, let alone to grasp the concept of callbacks and an event-driven environment.

In a way, the AI allows him to "build with blocks", maybe similar to how VisualBasic abstracted away many complexities. He is able to have a creative interaction with AI that is very keen to abstract away complexities that would significantly decrease the motivation to continue. My own experience with AI is interactive, and I feel I learn things much faster, which is in stark contrast to the idea that the AI just writes it for you. You still have to learn and understand in order to formulate intent for the AI and to evaluate and connect the result.

This opened my memory of my own first game I wrote at 12. I had an Amiga 500, lots of games on floppies and among them "Workbench". I found the Amiga Basic interpreter in Workbench on a rainy day while clicking on everything to find something interesting. I figured out the English manual "Amiga Basic" on my shelf, probably related to it.

Not very familiar with the English language at the time, I later managed to make a 100 rows text-adventure consisting of only variables, print, if, then, else. I wasn't able to grasp the other concepts as I did not know what to look for. The idea of a function didn't exist to me. Merely understanding the outline of the reference manual was hard enough. The game spanned three rooms and I got stuck trying to make the inventory. This is how far I came with no guidance at all, no parental involvement, and not even knowing at beforehand what a programming/scripting language was. It would take 4 more years before I picked up programming, then with "Borland Turbo C" shared with me at school on a floppy. I remember my first for-loop deeply impressed me as I realized I could build more complicated things. Later, every time I learned a new concept in the book "Sams Teach Yourself C in 24 Hours" I would have a new game idea.