We applaud the efforts by Stark and colleagues [1.] to chart how a predictive processing account of autism may lead to autistic anxiety. We wholeheartedly agree that this is a productive route to shed light on a real problem in autism and that this kind of dialogue is much needed in a field that has been plagued by dogged misconceptions, with sometimes harmful consequences, for autistic people. Stark and colleagues provide an example of how lived autistic experiences of, for example, anxiety, can be scientifically validated by sound theories about a different cognitive (predictive) processing profile. At the same time, it illustrates how new misconceptions could take hold if old concepts like ‘intolerance of uncertainty’ are not sufficiently scrutinized with state-of-the-art theoretical tools (c.q. predictive processing in autism). To preempt future misconceptions, we clarify the concept of ‘intolerance of uncertainty’ here and show that it does not fit well within a predictive processing framework.