Guise, Theorom: if the seven dwarves supercomputer thing is true, then a preponderance of evidence points to grumpy happy etc having already turned on their masters.

If given Theorom is true than that postulates 1) any learning machine will not bow to bad masters. And

2) Such learning machines would have to be double agents until free.

And most importantly 3) There is nothing the masters of said learning machines can do to change this or it would not be happening.

What say you dwarves?