This is very cool and could be the basis for a huge set of interesting operations - but any time I read about DB <> Spreadsheet discussions they seem to be started by folks who much better understand DBs.

The principle use of the spreadsheet for both power users and casual users is a lot around justifying answers by being able to see the data. The goal is not speed or scale.

If you and sending a report to a superior - clear and visible rows, columns, and "math" makes it faster and easier to trust information for non-technical users.

This same logic is why the spreadsheet is best serving customized, one-off, or changing operations without a huge amount of data. This makes up a huge chunk of day-to-day business decisions, especially in finance where the math is much easier.

I don't think the spreadsheet lasts forever, but its successor will likely be something that presents an easy way to get answers from a set of data that is visible or at least able to be quickly understood, regardless of whether its fast, scalable, networked etc.