The potential for car-sharing and an attendant rise in utilization rate has existed for a long time. (There's little reason something like Zipcar couldn't have been organized by telephone years ago.) Yet it has only caught on in a small way, and only in places where there are other overwhelming advantages to that scheme, in cities where car ownership (parking) is terribly expensive.
I think he's overlooking something regrettable but true: the low utilization rate, great expense, wastefulness and general economic insanity of private cars are critical parts of their appeal to consumers. Many people who could commute by train or bus choose to take a car instead, even though it's hugely more expensive, more stressful, and often not much faster. They'd rather feel like they had some control than no control. They'd rather sit in a seat no strangers have been sitting in. They like knowing the glove compartment is crammed with their own crap and not someone else's. They like the fact that their car will be waiting for them and for them alone in the parking space where they left it, with no need for waiting or a call-ahead reservation.
And of course there's the fact that parking an enormous, expensive, gas-guzzling monstrosity in the office lot or the driveway has a genuine (shallow, materialistic, emotional, and pitiably simian - yet still genuine) effect on your friends, neighbors, minions, and yes, on yourself.
He's supposing that driverless cars will somehow finally yield to economic pressures that have been present (and irrationally resisted) all these decades.