Ask HN: Are there good reasons not to deploy Machine Learning models?
It's been reported that as much as 80-90% of machine learning models that are meant to be deployed never get there. Is it more intentional or just a problem with the fragmented nature of the AI industry? Not every use case requires deployment, sometimes you just save your results to a file/database, and/or you write a report/paper/presentation etc. Yea, and I understand researchers may just be training models but don't intend to move into production or deploy once they fine tune them, especially in an education environment or when using them as a basis for a research paper. It's not just researchers, there's actual businesses using ML that never gets deployed in any way, and they don't need it to be deployed.