I present a subjective list of what I think are the most serious problems in the modelling of AGB stars. Because AGB stars represent the last phase of stellar evolution, they suffer from the accumulation of the effects of uncertainties in all the earlier phases. The complexity of AGB evolution adds further uncertainties specific to the evolutionary behaviour of those stars. Most of the problems are associated with mixing, specifically the boundaries of mixed regions. The nature of the "extra-mixing" remains a mystery, let alone how to model it reliably. Other problems are briefly mentioned and I finish with some hopes of making progress in the future.