With microservices, shared DB approach is bad because it leads to coupling at the DB level and prevents a service owner to confidently release the new version of the service as the service owner does not know how the data tables of their service is used by other services and if any schema change will break the other service or not. Plus few other issues.

At the same time - every architecture is context based and cannot be generalized and hence, there could be a use case where sharing a DB can help solve the situation in hand.