Hacker News Reader

Comment by dumbfoundded | Hacker News Reader

Uber really has no advantage to self-driving. A customer network doesn't mean much if you cannot finance actually building the cars. They also don't any training data to give them a tech advantage.

It'd make way more sense for an existing car company to build the features into the car and let the owner rent out the car when they don't need it. (Telsa, GM/Cruise).

Another option is to simply build the best technology and use ungodly piles of money to finance building the cars. Only Google or Apple could really do this.

View on HN

I don't think see the point of owning a car, if it's used by other people for most the day. I think there is going to be a big company that owns most cars.

Uber could raise the capital for it on the stock market. The problem that I see is that Uber doesn't have the technology.

In a world where self driving cars are a commodity, Uber's brand could be an advantage. But in reality, by the time Uber gets access to the technology the market has already been taken by others.

This is a key bit of psychology that is often overlooked. People who get Ubers exclusively think that a car is a metal box on wheels that goes from A to B. But most people who own cars see them as a self-propelled room of their apartment or house, a personal space in other words.

Like if you have a kid and drive your kid to places in your car then you probably leave loads of kid stuff in your car, and if the kid spill something you probably don’t worry too much about cleaning it to pristine state immediately. But imagine after every school run or activity or playdate you had to valet your car immaculate so strangers could use it. No one will do this.

I've thought a lot about this. I think, if the price is right, it could work for families in lieu of a second car. In a "traditional" family there's two vehicles. One is a commuter car and the other is a family-mobile (Van/SUV) which holds all the stuff. In this case, the commuter car could theoretically be replaced by a car service in many cases.

The other problem is that people would have to stop seeing cars as a status symbol. A not insignificant portion of women I know think it's hot if a guy has a cool/fast/whatever car. That will be harder to change.

At least in my whereabouts all (solo or both) cars have same role. One for mom and another for dad. Likely one is coming in later to drop kid off and other is leaving job early to pick him up. Or maybe multiple kids go to different institutions. There's no clear separation between kids-ok and no-kids car in family.

IMO the fantasy of car sharing is naive and I'd even say it's counterproductive for the supposed goals. Just a few bits to consider:

- empty cars are going to drive a lot of junk miles. More resources to be burned, more wear on cars.

- car-sharing companies will try to keep brand-new fleets. Reliability, image and whatnot. What are we going to do with used-but-drivable cars?

All of these services and capabilities make a difference at the margins--which a second household car often is. But I suspect that a lot of the "people won't own cars" refrain comes from people who either don't own a car or live in a city and don't use it much.

In a lot of couples, people still own their "own" cars, keep their stuff in them, customize them for specific uses (am I going to get an Uber with kayak cradles?), and so forth. Even if the finances worked out, there's no way that I'm dumping my vehicle for a generic shared car.

Owning a car is a luxury, and part of the psychology of a luxury good is its exclusivity. Obviously, a “production” sports car limited to 25 examples is one kind of exclusivity, but a car that is exclusively for your personal use is another kind of luxury.

And with that exclusivity comes the right to personalize your experience, from features to things for your kids to bumper stickers to blasting your music rather than having to ask the driver for another station or listen to your music on air pods.

The luxury of owning your own car may not appeal to everyone, but I do not see car shares and ride shares replacing cars for everyone at this time, it would take a major cultural shift and then several generations to age out of the market and die off before it ceases to be an economic factor.

I'm not sure even the finances would work out.

So you have an asset--your car--that let's just assume you're fine with it being something economical to drive and you don't need to keep a bunch of personal stuff in it. Now you rent it out.

So you're in the single vehicle fleet management business. Your asset depreciates mostly on a per mile basis. (Not entirely but a good first approximation.) Say 50 cents a mile cost (including fuel). EV eventually brings this down but who knows what incremental hardware will be required for autonomy.

So two possible outcomes here.

One is that reimbursement costs drop to near operating cost. Some people will want to borrow from their future selves or will figure that their costs are enough below the average to make it work. In any case, it's not a great deal for the car owner.

The other is that reimbursement costs somehow embed a nice profit for the car owner. In this case, why on earth wouldn't the sort of companies that manage car fleets today get into the business, especially given that they can do things like maintenance less expensively than an individual.

Uber has an advantage bc the transition to fully autonomous self-driving will take years to decades to realize, not only on a technological level but on a political level. In the meantime we’ll introduce self-driving more and more with humans-in-the-look, and uber is in a great position to take advantage of that.

I agree. Self-driving cars will take a really long time to emerge. The progress will be gradual. According to George Holtz of Comma ai, great lane-keeping is the first major milestone.

The problem with your logic is the connection to Uber. Uber doesn't control the cars their drivers use. Sure they can set restrictions but they never collect actual data that is useful for self-driving. The only data Uber collects is when and where people want to travel. Google & Apple both have this dataset (probably superior) through their map apps.

The real data gathering will come in the way Tesla does. Gather data from real-time drivers by putting the necessary sensors in cars you sell.

Maybe Uber could force their drivers to install kits but I doubt it. Self-driving cars will emerge over time as new features in cars you buy. This is already happening (ie brake assist). You'll then just get updates to your car when new features become available.

There's a pretty good talk about the future self-driving cars here: https://www.youtube.com/watch?v=iwcYp-XT7UI

No, it makes more sense for Uber to do so because the solution they would build “could” be an add on to different cars giving anyone with such car the ability to add it to Uber’s fleet. When BMW creates their solution, it will “probably” not work on an Audi or Ford.

That's just not how self-driving cars will emerge. There's not going to be level-5 for a really long time. Even longer to get the laws into place to manage it. The progress will be gradual.

What's going to happen and is already happening is cars will offer certain self-driving features. Things like lane-keeping, brake assist, highway driving. Car companies will also put in all the necessary sensors to collect training data. Then provide over the air updates. Tesla & GM(Cruise) are already doing this.

There's not really a way for Uber to get to level-5 first and at that point, why build a kit that integrates with your old Honda to give it lane-keeping?

The only dataset Uber has is a general sense of where and when people want to travel. Google & Apple probably have much better data through their map apps. They both also have an ungodly amount of money to both acquire the best researchers and create & distribute these kits.

In my opinion, self-driving cars are a race between auto companies and these cash-flush tech giants. Uber/Lyft/Didi all don't really have a chance.

>self-driving cars are a race between auto companies and these cash-flush tech giants.

The auto companies and their suppliers like Bosch. And, if Apple, for example, were to come up with a really good system for autonomous driving or a subset thereof, don't you think they would be happy to sell it to automakers who already buy a huge number of component systems. After all, they already do with CarPlay. (Which is better than any automaker system I've used.)

Reply on news.ycombinator.com