If you can do processing outside of the database you don't have a lot, that's basically the definition of a lot of data. Don't be sorry, try to not have more data.
That's an odd rule of thumb. You going to do some processing outside of the DB, e.g. transformations, views. Then you have the event streaming and aggregations world. Then different type of DB. Distributed DBs where you could do processing on each node in code for example.
I think we agree. Transformations and views bring code to data (if you are willing to consider them code, which I am). Then when you say 'Distributed DBs where you could do processing on each node in code for example' it also sounds like exactly what I have in mind - code brought to data, not data to code.
Once you have data you move code to data and not data to code. It isn't a decision you get to make. You only decide how you do it.