Yeah, that's my only real issue. Lots of people have written artifact quality pathing implementations. He should probably break that part out to a dedicated core or three, or provide enough dox to allow others to implement it.path finding
Also, I'm pretty sure some of you guys are underestimating the complexity of this simulation by an order of magnitude or three.
e: artifact quality = BAD in this context. expensive in time and resources and good for nothing but attracting unwanted (usually negative) attention