In August 2016, Ford announced its plan to offer a fully-autonomous, level 4 vehicle (no driver required) in 2021. On Monday, at the Further with Ford conference, CEO Mark Fields gave some more information about Ford’s plans. Henry Blodget, co-founder, CEO, and editor-in-chief of Business Insider, asked Fields for details on the “where, when, and how” these cars will be available. Here’s their conversation, lightly edited for clarity.

When can I buy it, and how much is it going to cost?

There’s lots of discussion around what is an autonomous vehicle. The Society of Automotive Engineers have five levels of autonomy. A level four vehicle is a vehicle where the passenger never has to be prepared to take control. Our intent is to have a level four vehicle, without the steering wheel, without a gas pedal or a brake pedal in 2021. It will be used for a ride-share or a ride-hailing service, so not initially available for, let’s say, personal purchase. Of course, the vehicle, because of the technology, will cost more, although that cost is coming down. That’s why it’s so compelling in a ride-hailing service, because when you take the driver out, that’s the biggest cost of the ride-hailing service. That’s probably the reason the majority of the ride- hailing services aren’t making any money, so it’s very compelling.

Where will I see level 4 Fords in 2021? And when can I actually buy one for the family?

It’s going to be in limited areas. For example, we want it in large cities, in dense areas. The reason for that is we need to launch them in what we call “geo-fenced areas.” And what that needs is mapping. All these areas need to be mapped. The greatest need will be in cities, because people want to get around and it’s tough to get around.

How much will the self-driving technology cost? How much does it add into the cost of the car?

Well, today it adds a good amount of cost. What you usually see when you’re developing new technology through cost fusion is two different things: Scale and improvements in the technology. We expect between now and 2021, we’ll make some progress, and then as the decade goes on, further progress. Probably around two years later, mid-decade, we will make vehicles available for people to purchase for themselves as opposed to just a ride-hailing service.

There’s a company in Silicon Valley that’s experimenting with self-driving technology and the cars are, shall we say, a design departure from the cars that we’ve thought of. The car that you will have in 2021, will it look ridiculous?

It will look very, very nice. When you’re a passenger in a vehicle, and you’re not focused on driving, think about the excitement that brings our designers. For instance, how the interior will be designed. Think about, well, you can imagine all the things you do or can do in a car. Think about all the things that you can do when you don’t have to be focused on driving? You can design lots of things, using the interior for a lot of different uses.

SEE: Ford plans to mass produce a ‘no driver required’ autonomous vehicle by 2021 (TechRepublic)

One of the things that I noticed this morning is that the drive went nearly perfectly. We waited at the stop sign and a pedestrian very safely crossed the road. Then we encountered an unplanned pedestrian, who said, “no you go first.” The car waited very quietly and didn’t do anything. I would describe it as a small hurdle that has not been solved yet. The car didn’t know who was going to win the “polite” contest. What other hurdles need to be solved before 2021?

I think there are a number of hurdles. Put it into four buckets: There’s the technology hurdle, there’s the regulatory hurdle, there’s the economic hurdle (in terms of bringing the cost down), and just the adoption hurdle of people getting used to this. There’s also the technology bucket, if I can call it that. You have to be able to develop technology for the rules of the road, and then you have to develop the intuition that’s not the rules of the road.

SEE: Autonomous driving levels 0 to 5: Understanding the differences (TechRepublic)

Here’s an example: You come to a four-way stop, and you happen to come to that four-way stop at the same time another car does. Today, I’m looking at the other car, seeing if it’s inching forward, or I look at the driver in the eyes and I see, okay, they’re going, so I’m going to let them go. We have to be able to anticipate that. Or, whether you show up in a construction area and a policeman or woman is there directing traffic. How do we develop for hand signals, in the face of that issue. That is the working promise, and that is why we’re so excited about that.

Part of why I was so sold by these vehicles was that I actually needed no convincing, because I’ve heard many of the companies that are working on this in Silicon Valley and then other car company executives say, “You don’t have to be secretive that cars are already way better drivers than human beings.” One executive confessed that to me–she’s a marketing executive for a very high-end car–she said: “This kind of creates a problem for us, because we’ve spent the last 50 years selling the ultimate driving machine and now we have to sell the ‘well, it will drive you, but you don’t actually have to drive.'” Does this create a marketing problem for you? Not that you’d say, “It’s been a bust, just kidding.”

SEE: Why Ford is shifting its focus from cars to ‘mobility’ (TechRepublic)

There is going to be a spectrum of drivers going forward. There’s going to be a spectrum of folks who are always going to want to drive and have that pleasure of driving. There’s also going to be a lot of folks that view driving as a chore, and they want to be freed from that. I think, from our standpoint, in our company we use the phrase “fun to drive.” That’s still very valid going forward. We also added “fun to ride.” So we’ll see how far that goes.

Let’s talk about Google, Apple, companies that are fiercely competitive, known for technology. You have a wonderful history of making and selling cars, but do you believe Silicon Valley will disrupt this?

The world has moved from a mindset of owning vehicles to a mindset of owning them and sharing them. That attitude is a really important part of how you approach things. We’re embracing our core business of designing and developing great cars, utilities, and trucks, and embracing these opportunities around the usage of our products, in terms of these mobility services.

SEE: Tesla’s Master Plan 2.0: AI experts, auto insiders, and Tesla customers weigh in (TechRepublic)

We have always said, “Listen, we will do things on our own, some things we will partner with, some other opportunities we will acquire,” just like we acquired a software company a couple of weeks ago to help us in the autonomous vehicle area. But if we embrace this, it’s not moving from an old business being a core business to a new business in a mobility area–it’s just moving to a bigger business. We’re really embracing this. Not viewing it as a threat.