There is a long tradition of research into autistic people's executive function skills. Yet, despite decades of research on EF in autism, the existing literature remains contradictory and 'confusing', with a large -- and unresolved -- discrepancy between small-to-moderate effect sizes demonstrated on laboratory-based EF tasks and large effect sizes on questionnaire-based everyday EF measures. We sought to understand this mismatch between 'lab and life' by inviting 12 autistic adolescents (12-19 years) and their mothers (n = 7) to convey their views and perspectives about their own, or their child's, EF skills as they transition to adulthood. We followed Braun and Clarke's method for reflexive thematic analysis using an inductive approach. Participants told us that their EF skills were highly variable, acutely dependent on the context in which they were deployed and potentially related to differences in the way that they process information more broadly. Participants' reports provided rare insights into their and their children's executive control -- insights that do not straightforwardly map onto traditional theoretical models of EF. Future work on EF needs to take seriously the perspectives and subjectivity of autistic people themselves, including by triangulating quantitative, objective assessments with qualitative, subjective reports in complementary (controlled, uncontrolled) settings.