Example of nonexistence of continuous nonlinear max-likelihood smoothing estimate

An exponential function of a Gauss-Markov process is used here as an example to show the nonexistence of a continuous joint max-likelihood smoothing estimate over the interval
T \doteq