I think a route like GET /tables/{tabledId}/records/csv would be nice to have in the Tulip API. This would save developers from having to implement this manually using _sequenceNumber ordering. It will also reduce the number of calls made to the API (especially on large tables) and could rely on (already existing?) performance optimization used when manually clicking the export button.
The need to get all data in a Tulip table to transfer it to Power BI or some other data warehouse/mart for BI is probably quite common.