Eric Schmidt is one of the more unfiltered executives in Alphabet's ranks — so if you're going to get a straight answer on when we can truly, legitimately have a fully self-driving car pick us up at our doorstep, you'd think the answer might come from him. But at the company's stockholder meeting today, he demurred.

Responding to an audience question about how close self-driving cars are to reality, Schmidt had this to say:

It's very hard to know. The consensus I think within the company is that it's some years, not decades, but it is very much dependent on regulation. And it also depends on where you are. It's obviously a great deal easier to do this in areas that, for example, have ample parking.

Most of this is terribly surprising: every company in Detroit, Silicon Valley, and elsewhere who is working on an autonomous project has mentioned that regulatory concerns are one of the biggest roadblocks to deploying the vehicles quickly. (And, not for nothing, the US Department of Transportation and NHTSA seem to be responding as quickly as bureaucracies can.)

But the "ample parking" thing is a little more notable. Google's Chris Urmson has said in the past that a big promise of self-driving fleets is reducing or eliminating wasted space in urban centers because parking lots and garages won't be needed as much there — an autonomous vehicle can drive off and park somewhere less interesting, assuming it isn't working the roads in ride-hailing duty 24 hours a day. But Schmidt seems to suggest that parking will be more important, at least for initial rollouts — and it goes without saying that the company's current test cities (Mountain View, Austin, Kirkland, and Phoenix) are more open and parking-friendly than, say, San Francisco or New York City.

Either way, Schmidt hopes the cars start being real in a matter of "years, not decades."