This article pushes for an ethics and feelings based redistribution of programming effort. That's bad. Because resources remain finite, we won't lose less data, we can only shuffle around what we lose. What we need to make this decision, is not feelings or vague ethical codes. We need hard data, especially in an age of big data and machine learning where the detailed pieces involved in making a deep statistical analysis may well be out of the realm of human comprehension.
For example, this article targets the evil gender binary. If only we would think more carefully about our choices. Meanwhile, clinics are scratching their heads trying to figure out how to map rates of ovarian cancer to "demisexuals", or how a local rise in "non-binary asexuals" might affect rates of testicular cancer, and what should be done about funding for screenings.
We lose data because we come from an era where space was limited, and cycles were counted. Sure we made some dumb mistakes like 2 digit years, but 2 digit years still work in a broad array of cases so we still use them all over the place today. They are sometimes good to use, sometimes bad, sometimes they're superfluous and unused anyway.
By all means rethink what fields should be compressed or expanded based upon business needs. But do it based on data and business use cases, not on feelings and someone's personal bias- even if that person promises their ethics are better than the old ones.