nothing was done, and much smaller numbers (8-36K) if there was a lockdown. And there were several other potential scenarios modeled in between these two "bookends." The paper includes a very handy sensitivity analysis showing their projection of deaths with various scenarios, including these.
I'm not going to defend the details of the paper, but yes, there has been an overreaction on CB today, and a gross mischaracterization of what happened. The model didn't change — the scenario being modeled changed.