The first draft of a discussion of how the Dickstein-Morales discrete choice estimator works, how to perform inference in that setting, and some practical pitfalls to be aware of is posted here. I will update in the future with a Monte Carlo study of coverage and rejection rates for various implementation choices.