Hacker News Reader

Comment by jcranberry | Hacker News Reader

I wonder how this works for 3D game engines? It would be positively insane if AI could automate one of the most complex and high skill requirement positions in the video game industry.

View on HN

What would that position be? You realize that this is about copying the mechanics of an existing game, right?

You could maybe adapt it to learn from hand-animated examples, which would certainly make some games easier to create. But I doubt that the effort required to specify all interactions in a 3D game by hand would be much less than just coding them.

What if you trained it on footage from the real world? I didn't see anything in the article that stated the AI needed any sort of feedback from interactive controls...

It doesn't need feedback from controls, but they have to be present as inputs during the learning process, otherwise you can't hook them up correctly. Learning real-world physics from video would be impressive in its own right, but alone it's not enough to create a game. It's also not necessary, since we can already simulate most physical phenomena; and much more efficiently than what a learning process is likely to produce at first.

In addition to the limitations stated by sibling response, you would also need to solve the problem of image classification before you could start on real world footage; the ai in the paper was given the videogame sprites in addition to the raw footage.

Interesting idea. As I understand it, you must give it a file with all the objects that will be shown in the game first. I think the general idea might be applicable to 3D but I think you'd need to give it all the objects that could appear as 3D models first.

Reply on news.ycombinator.com