[Mvhsz]

In reference to the cmip model, the article says:

> There are, however, model biases toward overestimated stability of the AMOC, both from tuning to the historic climate record, poor representation of the deep water formation, salinity and glacial runoff.

The article goes in to say why they disagree with the cmip assessment, and rather believe it could happen “around mid century” of the current century.