It's a mineral. It's electronic properties are almost completely random. The only characteristic of it that remains the same is that it exerts more electricity than you put in and it cools down in electric current.
It's random properties can only range in these few: Only conducting on it's surface. Suddenly changing electronic resistance. Swirling electricity. Sending electricity in a straight line. Only conducting to random conductors. Stops conducting.
Even the slightest interference will cause it to change properties. But it can be controlled by some elements.
If random properties are tricky to code, I suggest the properties be determined by whatever the conductor sparked next to it first is (i.e. if sparked by METL, it sends electricity in a straight line; if sparked by nscn it stops conducting...just a suggestion. On the other hand it sounds as if bismuth telluride could end up as a sort of swch (pscn on, nscn off) but combining several electronic functions in one element (i.e. ntct turns on state 2, ptct turns off state 2; metl turns on state 3, inwr turns off state 3)...any thoughts?
I have a piece in my collection. If you dont care what its electrical properties are, it looks cool. I wonder you could grow it by flowing water over it?