There is a map-territory problem here.

There is some underlying reality to what autism is, even if we do not have a good understanding of it; and even if turns out to be multiple unrelated things that happen to have similar symptoms.

Of the people with those actual conditions, it seems entirely plausible that some will not be hindered.

The authors of the DSM-V needed to create a diagnostic criteria for a condition that they do not understand, and for which no objective test is known. Further, their objective was designing something useful in a clinical setting. Giving those constraints, saying "if it is not a problem, we don't care about it" is entirely reasonable; despite not being reflective of the underlying reality.