This technique, also called āopaque recurrence,ā will likely make the modelās chain of thought more difficult to monitor ā and that has AI safety experts rattled. While Astraās use of the technique is reportedly limited, its emergence has still raised significant concerns among AI safety experts.
