I don't think that a current limitation is an indicator of future success. I think a more likely future is one governed by diminishing returns, and prohibitively large computation.