Comment by andrewla

Is the plan to expand from weather forecasting into climate simulation? Given the complexity of the finding initial conditions on the earth, a non-physical (or only implicitly-physical) model seems like it could offer a very promising alternative to physical models. The existing physical models, while often grossly correct (in terms of averages), suffer from producing unphysical configurations on a local basis.