AI autopilot patch object connection
Auto evolving patch mode
Now we are ready for this, i'm thinking about a create random patch command/button, in order to auto add some random objects, connect them in a random fashion too, as a starting point, and then use the first 100 "corrections" ( as corrections i mean disconnects, reconnects, delete objects and add new objects ) by the user to feed some sort of neural network to learn generate better patches over time.
And in some way have a shared train model that is built from every user interaction with this feature
Thoughts?
What should we use for the neural network and the training model?
I have zero knowledge with AI... Idea: a "copy objects as string" and "paste objects from string" functions could help with such scenarios ? (and also for the new chat!)
I think it's a good idea, it shure helps with the chat, and probably will help with the AI idea, a toString(object) and toObject(string) serialization functions will be useful