Mark Broadbent saves some money via pool auto-pausing:
This capability is neither earth shatteringly new nor unexpected, and something that Databricks has provided for some time. Of the two Data Exploration & Data Warehousing Pool types, Synapse Serverless Pool (otherwise know as the built-in Pool) by its very definition does not incur compute charges when it is not running.
Therefore this leaves us with only dedicated SQL Pool to worry about and this is where our problems begin.
Click through for the scripts to pause and resume a dedicated SQL pool, and Mark promises a part 2 in which we see the automation.