Normally, people think of genetic conditions and hereditary conditions as identical. But, this isn't necessarily so. There are two main ways that someone can end up with genes that do not come from their parents.
First, there can be a mutation in a germline cell (i.e. a sperm or egg) that is not present in the rest of the genome of the parent. This is the mechanism suspected in almost all congenital conditions that are associated with advanced paternal age (i.e. with old fathers).
Second, a retrovirus can infect a person and change their genome during their life. These kinds of viruses are rare, but not unknown and their existence is fundamental to almost all proposed gene therapies, something done routinely in lab rats but only a few times on a therapeutic clinical basis in human subjects.
The distinction is important in interpreting a recent study of autistic twins in in California and receiving developmental disability assistance from the state that purport to show about 55% of autism is attributable to shared environment, rather than to hereditary causes or non-shared environment. Previous studies had suggested that autism was 90%+ genetic.
If there really is a large shared environment effect, it is probably a neo-natal environmental exposure situation, perhaps, for example, due mostly to pregnant women taking SSRI drugs. But, genetic factors like first generation germline cell mutations attributable to advanced paternal age and possibly also to the environmental exposures that the father has received would look mostly like a shared environment effect in the simple heredity analysis used in twin studies.
This method of analysis ignores first generation germline mutations, which is often a sensible thing to do, but is probably not appropriate in the case of conditions where there is an epidemiology that shows a strong advanced paternal age effect. Most serious dominant gene developmental disorders are probably predominantly due to first generation germline mutations, so excluding that possibility in a twin study analyzing autism causation is probably not reasonable.
The emerging consensus model of autism causation sees this syndrome as being caused by mutations in any of a very large number of genes (perhaps hundreds) that must all be perfectly in harmony to carry out the part of brain function that is atypical in autism, but can be mitigated if one or more "protective" genes are present (possibly an X-chromosome linked gene that could account for the differing rates of autism in boys and girls, since boys are less likely to have at least one copy of the protective X-linked gene).
In this model, most autism cases arise from first generation dominant gene germline mutations, and a minority of cases arise from inheritance from a parent who may be a carrier due to the presence of a protective gene that silences or mitigates the effect of the germline mutation that person received. The relative number of inherited and first generation mutations can be inferred by the extent to which advanced paternal age is a risk factor in autism.