Update quickstart to lower header (hotfix) (#117)

Author: jfomhover

docs/quickstart.md

@@ -4,10 +4,6 @@ In this tutorial, you will:
 * Provision a fully functional environment in your own Azure subscription
 * Run a sample federated learning pipeline in Azure ML
 
-:warning: **IMPORTANT** :warning: This setup is intended only for demo purposes.
-The data is still accessible by the a user of your subscription when opening the storage accounts,
-and data exfiltration is easy.
-
 ## Prerequisites
 
 To enjoy this quickstart, you will need to:
@@ -56,8 +52,8 @@ We will provision:
 
     > Notes:
       > - If someone already provisioned a demo with the same name in your subscription, change `demoBaseName` parameter to a unique value.
-
-    > :warning: EXPERIMENTAL :warning: alternatively, you can try provisioning a sandbox where the silos storages are kept eyes-off by a private service endpoint, accessible only by the silo compute through a vnet. To try it out, use template file `mlops/bicep/vnet_publicip_sandbox_setup.bicep` instead. All the code samples below remains the same. Please check the header of that bicep file to understand its capabilities and limitations.
+      > - :warning: **IMPORTANT** :warning: This setup is intended only for demo purposes. The data is still accessible by the a user of your subscription when opening the storage accounts, and data exfiltration is possible.
+      > - :warning: EXPERIMENTAL :warning: alternatively, you can try provisioning a sandbox where the silos storages are kept eyes-off by a private service endpoint, accessible only by the silo compute through a vnet. To try it out, use template file `mlops/bicep/vnet_publicip_sandbox_setup.bicep` instead. All the code samples below remains the same. Please check the header of that bicep file to understand its capabilities and limitations.
 
 ## Launch the demo experiment