Now, this is not a judgment for or against SQL, given there are plenty of users of Hibernate, it has clearly worked for some teams. Equally, given people still write SQL, I am sure it has worked for many people.
Fundamentally there are two questions to my post:
1 - how much could have been saved in this scenario by a better team
2 - how often is it that the right solution now becomes the wrong solution later?
If they had chosen to go with plain SQL, maybe the project would be finished successfully by year two.