My biggest practical issues are around a few libraries with really restrictive licensing without good floss alternatives. At least short of the effort to hire a

My biggest practical issues are around a few libraries with really restrictive licensing without good floss alternatives. At least short of the effort to hire and create appropriate ML models in house, which isn't a proficiency we currently have.