> Is this insane? No, in fact I belive some variant of this approach is what will eventually lead to AGI. (you can probably do a similar type of solution but a

> Is this insane? No, in fact I belive some variant of this approach is what will eventually lead to AGI. (you can probably do a similar type of solution but approaching from the direction of neural networks where you can use backprop)