How would you say self-driving vehicles fit into this? Supposedly they're rolling out in ~3-5 years, and with self-driving vehicles, you no longer need the middle man managing a fleet of contractors.
EDIT: I'm willing to me a Long Bet [1] with anyone regarding fully autonomous self driving vehicles in regular use in 6 years.
~3-5 years is way too soon. The technology might be ready but the governments (think regulations, traffic laws, car insurance companies, car manufacturers, etc etc etc ) in addition to the public (which is not ready either). Is going to take at least 10+ years for it to resemble something like you imagine it.
I agree that 3-5 years is a bit optimistic but I also bet that when it does happen it will transition far faster than we can imagine. You have a really nice set of aligned incentives here:
- As a 'driver' I get a car that let's me free up large chunks of my time
- As a company like Uber the driver 'input' to my model is now far more consistent, predictable, and inexpensive
- As an insurance company I can start undercutting my competition by charging less to insure this substantially safer means of transportation while still pocketing a nice margin
- As a car manufacturer I now have an extremely compelling reason to sell cars to people that would otherwise wait between 4-6 years in between car purchases
- As a government I can (if I'm smart) try and pass laws that will result in drastic reductions in accidents, deaths, etc... I'll also likely be bringing in a lot of business and money by being on the leading edge of this (less traffic and congestion, new opportunities for businesses that leverage these cars, etc...). Whatever revenue I lose in traffic tickets I can make up for in other ways (like a reduced need for traffic enforcement).
As soon as any of the above are able to make/save money this is going to snowball.
The lobbying power and influence of the teamsters and taxi unions is probably the single largest hurdle but if self driving cars are as safe and cheap as it appears they will be this will be a speed bump, not a roadblock.
I'm not so sure the future is as bright as you make it out for all those groups. At the very least this change is going to bring with it a mountain of uncertainty.
As a driver, you probably lose your job. Not with the 3-5 year out model, but the one that comes 10 years after that. Even with the 3-5 year model, suddenly the competition for your job becomes extremely tight and starts to focus on things like people skills and entertaining, since the automated car will be doing most of the "heavy lifting" most of the time in a way that is automated into a commodity.
As Uber, you save a lot of money but you also lose a lot of moat, it becomes much easier/cheaper for other companies to suddenly field a fleet of cars.
As a car insurance company things look really iffy for you, again not so much in the 3-5 year timeframe but a bit further out when the human is completely unnecessary since at that point the insurance will be more about insuring the companies making the cars and car software and less about personal insurance, which means there is still significant money to be made, but probably far less than now (if the safety of automated cars turns out to be as great as it should) and by far fewer players.
As a car manufacturer, you'll do well at the beginning but the eventual model of a networked driverless transportation system should require far fewer people to own fewer cars and still live as if they do own cars, which is a long term problem for you when you've built a business optimized to sell cars to broad market consumers on a 5-ish year cycle.
As a government you should, in an ideal world, be in pretty good shape for the reasons you outlined, but good luck dealing with the lobbyists from industries this may "disrupt".
I do think driverless cars are quite a bit closer than most people think, for a lot of reasons, I just think they are going to upset a lot of applecarts on the way in, which isn't necessarily a bad thing unless you're an applecart vendor.
The technology is not nearly ready. Not in a city. It's like flying drone delivery. Easy to demonstrate in a single controlled setting, but nearly impossible to actually implement in the real world.
Actually landing UAVs is far harder than automatically parking a car.
The driving vs flying after take off is what is easier. Landing and take off.. nope.
Note that drone could mean a car in this case. UAV being an unmanned aerial vehicule.
With a modest infrastructure of marked landing pads on roofs, I think UAV landings are easier than automated car parking. If you're expecting delivery UAVs to fly under the tree line along sidewalks and yards to land directly on a residential porch, then yes, that's extremely difficult. But I highly doubt that this is a remotely feasible plan within a decade or so.
It's hard for me to even imagine how a self-driving car could manage somewhere like large swaths of Manhattan where a certain amount of, umm, aggression is needed to make any forward progress. Though it will likely happen some day (probably multiple decades from now).
As it is, I'd even settle for voice control on my phone that could reliably understand me.
> a certain amount of, umm, aggression is needed to make any forward progress
They added a slight "aggression" back in 2011, if other cars aren't letting it out it starts pushing forwards a bit:
> Sometimes, however, the car has to be more "aggressive." When going through a four-way intersection, for example, it yields to other vehicles based on road rules; but if other cars don't reciprocate, it advances a bit to show to the other drivers its intention. Without programming that kind of behavior, Urmson said, it would be impossible for the robot car to drive in the real world.
I really don't understand why people think small behavioural differences like that require decades of work. Things that humans find difficult and stressful are likely to be the things that are significantly easier when you have a reaction time measured in milliseconds and full 360 degree vision.
I hope you're right but the very uneven progress of AI generally--especially when it involves real world interfaces--makes me skeptical. As I wrote elsewhere, even voice recognition is very much still a mixed bag.
But voice recognition is something we take many years to learn, even when are brains are highly plastic. I find it hard to understand people when they speak often. We disambiguate heavily based on context and expectation, very complex issues.
Oh, I fully agree. The fact that it's been the subject of so much research and remains only very incompletely solved demonstrates just how difficult a problem it is. I'm just saying that navigating a potentially dangerous machine through a physical world populated with erratic human beings is also very complex and difficult--albeit probably far less so in some contexts (limited access highways in good weather) than others (Manhattan, Boston).
Self-driving vehicles are the final commoditization of the rides-for-hire business. Personally, I'd be very skeptical about the 3-5 year time frame, but clearly they're coming at some point. When they do, I can't see any rides-for-hire business having gross margins higher than 5% (though the size of the rides-for-hire market may well expand greatly). Barrier for entry is basically nonexistent.
Scale and perks will drive that 5% up, I imagine. Sure, Joe Average can buy a Tesla Model A (for automatic, obviously) and have it operate as a AI-Uber Driver, but Uber can buy a fleet of thousands, allowing them better pricing, better access to charging stations, better data about which rides are most valuable and where to position their cars. With a bit better margins comes things like bottled water in the cars. This differentiates them further, etc.
It's clearly Uber's plan to replace all their drivers with self-driving cars in the future - I think they've even said so themselves. But even six years is a long time for a VC-funded business.
And what does "regular use" mean? That seems like the critical component of a bet like this.
"Regular use" is defined as you being able to hail one to take you at least 30 miles for ~$50, from either a mobile or webapp.
If anyone can buy a self-driving vehicle, Uber and Lyft are aberrations, as anyone (even FedEx and UPS) could provide self-driving vehicle livery services. Its no longer a marketplace business; its a fleet management business (and both UPS and FedEx are extremely capable at managing capital expenditures, fleet operations, and system capacity).
I think we'll see at least two main car providers in major cities plus some niche operators, then one major network that allows you to buy your own car to be committed to a pool as an investment.
Do you think we'll see public transit orgs (for example, PACE/CTA in Chicago and its suburbs) purchase their own self-driving vehicles to provide transit services?
This isn't just a city issue, but a country issue too. Vulnerability to both economic sanctions and espionage should be real concerns. You can take a look at the companies operating in Israel for an example of a market this would be an immediate concern. Likewise for companies operating in Washington DC. There are politicians and lobbyists regularly using Uber. Would this DC audience want to use Yongche, Kuaidi, or another Chinese service?
Another interesting consideration is who is building and selling these automated vehicles. Hypothetically a future Google building vehicles could acquire Uber and now they own the entire value chain.
A company such as Google could earn a substantial piece of the rider's revenue from advertising which would make it hard for competitors much like Yahoo & Bing under earned Google on search inventory. At that point it becomes very hard for competitors. Does ad based earnings sound unreasonable? Consider that companies, such as Delta, already are bidding their organic listings down on Google. They rank #1 for their name yet still buy the inventory. You could jump in the car and say KFC -- and KFC pays Google for the ride. Or you say "fried chicken" and Google gets even more money to take you to the nearest KFC (or may be Popeyes in that case.)
Secondly, consider additional activities in the vehicle. A car build to drive itself is going to be configured differently than a manned vehicle. May be a Facebook car arrives and you strap on the complimentary Oculus headset.
I don't know what it is, but don't discount the ability of a self driving car service to substantially differentiate itself beyond competitors in the future.
Any of these options will make it extremely difficult for public transit to compete. Certainly cabs are done without gifted monopolies. (Also worth considering if a self-driving cab company can retain any political power without having a fleet of human drivers)
Whether or not Uber is overvalued, they are taking the money. Probably the better choice. The US stock market keeps hitting highs, at some point there will be a drop, or may be a crises with junk debt (very low yields just like 07) and it is going to be a lot more difficult to raise money over night.
Someone will still need to manage the fleet of self driving cars. Who is going to fill them with gas, do routine maintenance, clean them after some drunk dude puked all over the back seat, etc?
If they're electric vehicles, they'll charge on their own based on closest charging station and the cost of electricity at the time they need their charge. Also, the only maintenance you'd require would be tire rotations/changes every ~5K miles, and a new battery every 150K-200K miles (Tesla automated the battery pack swap, I'm sure they could automate pulling four wheels off and bolting four new ones back on).
I agree someone is going to need to clean them out though.
Reliable wireless inductive charging for cell phones doesn't even exist. I imagine an inductive charger for a car would function very well as a harddrive eraser as well, especially in 'inductive road' form.
There will be significant power losses (10-20%) and alignment issues.
One thing I really like about the whole inductive road concept is that it totally removes the need for a battery, one of the most expensive parts of an electric car.
You'd basically have a full-size slot car (if you remember those) track (but without sliding contacts).
While high-capacity electric vehicles like a Tesla wouldn't be able to wirelessly charge, it should be trivial to build a physical dock they can connect to with no human intervention (similar to how SpaceX designed the DragonEye mating connector for autonomous mating to the ISS).
I'm well aware. I was referring to public charging infrastructure using wireless charging. This is almost non-existent due to the lack of compatible vehicles today. I understand that some electric buses are already using wireless charging in major trials.
I think the 'self-driving' feature of the car will be more like an autopilot - you'd still need a human at the wheel to occasionally override the system.
Hence, I see the self-driving car to be more of a competition to short-term rentals like Zipcar. In situations where you don't want to drive (have to work/talk on phone; after a you've had a few drinks etc), you'd still need a driver.
I think we will see self driving cars in regular use that do not require an attentive driver on the freeway within 6 years. I think self driving cars will put Uber drivers out of business in 10 to 20 years.
Even if the technology becomes publicly usable (stretching it), there are bigger obstacles - government regulation and fear mongering - that make me think it will not happen in this timeframe.
Google is closing in on over a million miles with no accidents, and have decided to focus on refining their intelligence for city driving. 6 years tops.
I'm super excited for self-driving cars - they're literally going to change the world in dramatic ways. It's sort of mind boggling once you start really thinking about the implications of it.
You may as well pose it as "what if we invented teleportation?" Many of the end results are going to be similar.
By which you mean they already exist today? My parents got a demo ride in one at a google event earlier this year. As long as the city involved is Mountain View, the vehicles already exist.
(Google's current effort relies on extremely detailed mapping of all the roads involved, but given the continued existence of the StreetView program that approach probably scales tolerably well.)
Mountain View doesn't really qualify as a "city" in the sense of city traffic. The roads are wide and not particularly densely packed. Not to downplay the accomplishment, but it's probably a long way from being able to handle San Francisco let alone New York or London.
EDIT: I'm willing to me a Long Bet [1] with anyone regarding fully autonomous self driving vehicles in regular use in 6 years.
[1] http://longbets.org/ ; http://longbets.org/rules/