Turns out when you optimize something for millennia, the truly optimal solution is not a simple grid. That picture is essentially a proof that engineering will always be needed. Because any final solution is complex. Even if it's parts are trivial.