The computational challenge of solving actually big statistical problems has
dominated the field of computational statistics for the last few years.  The
challenge is that it seems difficult to steal techniques from other fields, and
this has required the development of a lot of new ideas.  In  this talk, I will 
outline some of the problems in the field, talk about the ones where I think
some constructive pilfering could be quite useful, and outline an Asynchronous
Distributed Gibbs sampler that has appropriated 30-odd years of techniques for
distributed numerical optimisation.