[karamazov]

The immediate next thing is expanding to Databricks, followed by BigQuery and other data warehouses.

In the long run we plan to optimize source code (e.g. C++).