EDIT2: It seems that with a combination of this method and another method, it's possible to analytically continue functions past natural boundaries, in a way that seems quite natural. For instance, the blue is the analytical continuation of the function (in red) $$ \sum_{n=0}^\infty x^{(n^2)} $$ which by Fabry gap theorem has a natural boundary on the unit circle. What I find especially interesting about this continuation is that at $x=-1$ the function has derivatives that are 0 at all orders, and the continuation converges to those derivatives while not becoming the constant function.