Much of this uncertainty stems from how we study autism. We tend to group children together based on features they share at a given point in time, and then assume that children within a particular group will share a similar developmental path. But children develop at different rates, and their individual developmental paths are rarely linear.
If you are a student of Glendenningism, this notion is one that can be applied to anyone: especially in public school we make broad, and often inaccurate, assumptions about people based on their age, their diagnoses, their behaviour, and their past experiences.
Life and learning and development are seldom linear, yet so many of the systems we’ve built assume they are.