4 ms·

at least in the case of reinforcement learning, no. just because you can simulate the problem doesn't mean you know how to optimally solve it - ex driving a car