It's important to realize that the type of transit rider that uses GO generally exists on an entirely different plane than those who use local transit.
Firstly, consider distance. For many GO riders, the distance just from their house to the nearest GO station is twice as long as the entire distance to work for a resident of Toronto. This has enormous implications on how people can get to the station. It is impossible to provide local transit to the GO station for the majority of GO riders.
Secondly, consider the alternatives. Recognizing that GO's primary function is to take people either downtown or across the 407, local transit is impractical and non existent for this type of trip. If GO ceased to exist, there are only two outcomes. The first is that many of the 150,000 daily riders would start driving. The second is that those who decide not to drive will quickly start looking for work outside of downtown Toronto. Local transit will not pick up the slack, except on the insignificant Richmond Hill line, where in 10 years the average station will only be 1.5 km from a subway station.
Thirdly, most GO stations are located within industrial railway lands. Those that aren't are often so far from the city that there'd be no hope of building any form of mixed use or high density development on top of the parking lot in the first place. Parking is the best possible, if not only viable use of land for probably at least 50% of GO stations.
Bottom line is that GO does not compete in any way whatsoever with local transit, except for the Richmond Hill line. In a best case scenario, the typical GO rider would carpool to their station, thereby reducing the number of parking spots by maybe 50%. Yes, distance based fare increases would be better, and yes, it would be ideal to offer non drivers a credit to their fare prices. And yes, there are a few stations that could be redeveloped. However, the best way of improving GO ridership is still to offer as many free parking spots as possible, while at the same time improving train frequencies.