I understand the author point, but I think this is over-complicating a database table while losing most of the features a database can give you. This is not so

I understand the author point, but I think this is over-complicating a database table while losing most of the features a database can give you. This is not some new concept, however. I stumbled upon this concept two years ago with some dude promoting a "Vault" architecture, where you use a single "notion.so" table to store all your data. You create views from this data to separate topics. You'll then be able to "centralize" all you notion stuff in a single file; all while being able to link any two topics or more together. What hit me is that I can export the notion table to CSV and then this can be fed into an AI pipeline that might be able to predict my tasks better (like code). Only problem was, a couple of months into this and the notion interface became completely unusable. This can be done with a regular database. Though the views/interfaces to interact are not that easy to create. I didn't find an alternative (I tried airtable too)