I think their point is that for all intents, the column IS not nullable. It's nullable as an artifact of making live schema migration easier, with no blocking/downtime. But as the data model is concerned, it should not be nullable.

Sure, if one just leaves a column nullable due to negligence, one should check the actual requirements and enforce them to make invalid states unrepresentable. The author still makes too strong of a claim that becomes detached from the migration aspect, insinuating that one can just empirically test the database to check whether this is the case, to which I disagree.