ModelingToolkitStandardLibrary and Optimization

I posted a workaround in the issue. See if that works and if it does then we should add it to the FAQ.

Make loss(p) with a vector of 2 things and put the constant inside of the loss function.