Can residual stresses in a material cause it to change its shape over time? Suppose a metal part is milled to a desired shape. Will the shape of this part change at all over time due to the residual stresses developed in the part from milling?
 A: Yes it will.
One of the people I shared  house with while doing my PhD had spent a year working for a company that made satellites (I forget which). At the time they were making the chassis for a satellite by milling out a solid block of metal, and they found that if they simply milled it and left it the chassis would creep enough to go out of tolerance. They had to use a series of millings then annealings to produce a final product that was stable.
According to my friend (I have no experience in this area) the milling was producing a high density of dislocations and it was the thermal relaxation of these dislocations causing the creeping. The effect was small, but a satellite is a piece of precision equipment so the motion was large enough to cause problems. Whether it would be a problem in less demanding applications like car engines I don't know.
A: John is right, and here are complicating factors that are significant in this context.
First of all, a chunk of metal that is rolled and pressed into rough shape will contain residual stresses locked into its microstructure that reflect its deformation history. Those stresses will be released in anisotropic ways when a piece is cut or milled off of the chunk, causing the part to warp out of shape. This is true even if the part was cast from the molten state and the only way to prevent this is to bake the chunk in an oven before machining it at a temperature between 1/3 and 1/2 of the melt temperature in degrees C; this is where solid-state diffusional transport (which is thermally-activated like ~e^kT)becomes significant. Because the process is thermally-activated, its rate will be extremely slow at room temperature; the objective of the bake is to accelerate the anneal so it can go to completion in a matter of hours instead of ten years. The part will emerge from the anneal stress-free, but warped. Then the chunk is finish-machined to its desired net shape, and will not warp further during the process.
Machinists who have to work metal without the benefit of an annealing oven know about this and will machine the part in a tedious series of light passes so as to track the warpage and adjust for it as the net shape is approached. 
