Lyft Self-Driving
lyft.com
lyft.com
http://www.kevinklinkenberg.com/blog/autonomous-vehicles-a-v... https://www.citylab.com/transportation/2017/12/what-elon-mus...
So I tend to not get excited about Lyft/Uber and autonomous vehicles when presented as a solution for transportation. Whether they will be successful companies is a separate question.
Self-driving, solar cars solve all problems of mass transit and are far cheaper, since they use existing roads.
Could you leave those tools at work/training if they had secure storage? Alternatively, could the training facility offer a set of training tools while you keep your personal set at home for training?
Public transportation is used by people from all walks of life. Not just office workers - cooks, nurses, artists, school teachers, police officers, even employees of trade companies... and every citizen benefits from reduced traffic, lower infrastructure costs, and better health.
Using car taxes to fund public transportation would not materially impact the carpenter. They'd just charge a little more. An absurdly high $1000/year tax spread across 50 clients is merely $20 per job - a more realistic $200/y tax would add $4 per job. Considering that public transportation increase average disposable income, the market can easily absorb that cost.
And so too can the office workers. To charge one group for the benefit of another, based solely on the manner of their employ, is the definition of unfair. This is what progressive taxation is for, not use taxes.
For businesses that require car usage, the tax is merely an operating cost. There is no impact on profit as operating costs are passed on to the consumer.
Of course, allocating that revenue is still subject to local voters' wishes.
Regardless, transit should pay for itself in the same proportion that roads pay for themselves. Transit has per-use or per-distance fees around using it, but are also subsidized by the government. Road travel has per-use or per-distance fees (tolls, gasoline taxes), and are also subsidized by the government. I'm sure we can continue to find a balance for all of those numbers so no one is disproportionately burdened to support their needed mode of transportation.
I agree that public transit should be improved. Transit should be free to the poorest people. It should also be accessible.
My point is just that existing mass transit is unpleasant or not viable for many people. Public transit in the form of solar SDCs has the advantages without the disadvantages.
They don't solve the arguably most important issue: road capacity. A bus that takes the same space on the road as 2-3 cars seats 30 people (and holds another 30 standing). Self-driving cars enable somewhat more efficient traffic patterns, but they will have a hard time overcoming the order-of-magnitude lead that buses have.
And that's not even mentioning subways, those don't impact roads at all and move staggering amounts of people
it doesn't matter how full the bus is when the road is empty, but the number of cars that a bus is taking off the road at peak road usage time is pretty close to the full capacity of the bus, in any bus system i've ever ridden.
Traveling with kids? Something people do with public transit and walking every day.
Promoting walking encourages healthy lifestyles; promoting driving (or just sitting in a car and not doing anything) encourages the opposite.
Elderly or disabled people take public transit all the time.
Short on time or carrying a bunch of groceries? Live out in the far suburbs? Then public transit isn’t for you. But it takes passengers who aren’t short on time off the roads, making your trip faster.
Edit: Cities with bad weather have adapted the cities to support public transit and walking (think covered, elevated walkways in Asia).
It’s not a zero sum game. Self driving cars will be great. Public transit is great (and can be great in more places 0if cities made more investment). Both can and should exist.
I'm a little bit worried about self driving cars being used for auto-detaining people and delivering them to police/terrorists. When you get into a sdc, you lose your ability to get off anywhere you want, it might lock up the doors and deliver you to the mafia or police for all you know.
Remember "auto-detention" I invented it here and now. :-)
I agree public transit is unequivocally great and important. I don't agree that most of our investment dollars should go to mass public transit over SDC public fleets.
NYC is an exception and can't be replicated in most places, and the lack of transit options are still a burden on the groups I mentioned.
Better public transit is also great, but that's also hard to deliver in low-density suburbia.
I agree that autonomous vehicles and Uber/Lift won't make travel times shorter. But they might help make the trip better because you can use the time better despite being in a car.
I do like trains, but they can't get everywhere. I take a bus to work everyday but they don't operate 24h a day and not frequently enough.
I'm the most excited about self driving buses. Even in predefined routes on special self driving lanes. Then its just a train carriage on a road :)
Agreed. Self-driving cars may be a sham "solution" to American's transportation woes....however self-driving busses should be considered as a very practical mass transit alternative to rail since the road infrastructure is already built. I've heard they are testing this out in Las Vegas:
https://blog.caranddriver.com/letting-it-ride-las-vegas-puts...
I live in a place where you don't get any of that that (unless you have $2M+ to spend on a house and land, which still doesn't fix the noise problem), and yet people still object to modest density increases and investment in public transit.
And in this case, I'd expect an increase in density to mean prices will drop, therefore enabling people to own more space at a lower price. This does mean more vertical building, so you lose private outdoor space, but that's really the only downside vs. the status quo.
It's an interesting to see all the players trying to have an out if someone withholds either the transportation network, or self driving tech.
Hubert Horan [1] has done a great job exposing Uber's failed business model in 11 part series on NakedCapitalism ("Can Uber Ever Deliver?"). Same goes for Lyft.
When it comes to self driving cars, why should Uber or Lyft have any massive advantage over other players and what makes us think it won't turn into something like an airline industry which has been unprofitable for most of their history?
It's possibly that self-driving cars won't be profitable, but it certainly won't be for the same reason.
There's reason to believe it'll be a winner-take-all market, because more data leads to lower crash rate which (in a world of rational consumers) will make them more popular, in a self-reinforcing cycle.
The U.S. Bureau of Labor Statistics reports that the average salary of a Commercial Pilot is $73,490 per year.
This seems low given the sheer number of hours of training involved.
But your last point is interesting. AlphaGo Zero suggests that for many situations, the return to more data levels off very quickly—after all, humans all learn to drive reasonably well with a certain amount of training. The early phases may have this self-reinforcing bit, and there may be an early dominant player, but we should expect that twenty years in anybody will be able to make a self-driving car that is as safe as modern air travel.
I’m not sure how well they will be able to diversify their product while keeping the economies of scale. If 10% of your customers is willing to pay for more leg space, can you get by with 10% of your fleet being larger cars, or does that affect the waiting times of that 10% too much, and do you need to make most of your cars larger?
Also, if you pay $1 extra per trip on your commute, that’s a nice smartphone each year.
Or, perhaps you have a small frame and the difference in leg room wouldn't impact you at all.
Also one man’s “reduced comfort” is another man’s “less luxury”. People make those choices all the times, e.g. when buying cars, train tickets, housing, or food.
How far off/difficult would it be for Tesla to create their own ride sharing app?
I mean both Lyft and Ubers "moat" seems to be their app which is not defensible against a market leader in autonomous cars that can simply switch on a new marketplace for its userbase and whole lot of other folks that can't afforda tesla but wants to be picked up and dropped off by one.
Would it be far fetched to see Tesla's market cap will combine Uber and Lyft on top of their existing market cap.
I mean it's a matter of time for Tesla...thanks to Uber's aggressive campaign efforts in spreading ride sharing, city management will be more than familiar and open to new innovators....plus Elon Musk's reputation (he might outsource some SpaceX to your city, you never know) riding him to the stratosphere....
Tesla's suburban HQ in the hills may be less appropriate (due to lack of density), or possibly more (easier to drive in suburbia vs the city), but at least one other manufacturer - Cruise, now owned by GM, is running internal trials, with hopes of opening it up soon.
They were offered more "efficient" machinery, which could manufacture more textiles than the old machines. "It will pay for itself" was the pitch. But they refused as more efficient machines meant they had to pass on the savings to the customers. With lower prices, machinery was not exactly going to pay for itself.
I guess that will be the case for Uber and Lyft too. Any efficiency with low enough barrier will mean no one is reaping profits rather passing it on to the customers and hence making less money.
Edit:
Found the link:
https://www.farnamstreetblog.com/a-lesson-on-worldly-wisdom/
I have included a link in my previous post where Charlie Munger talked about this.
That is the problem with commodity business because consumers dont care where there corn or cloth material comes from. Consumers will go for the cheapest option.
Google is mining millions of captchas for their self driving cars, they could publish their algorithms for it too, but unless you have access to all that beautiful training set, it wouldn't be much of use.
I have no idea how lyft or uber is going to get that much data, maybe potentially buy paying drivers more to stick sensors on their cars to collect insane amounts of data? But it's gonna be a tricky thing to get right
Detailed cost data from studies of traditional operators in Chicago, San Francisco and Denver showed that 58 cents of every gross passenger dollar (fares plus tips) went to driver take home pay and benefits, 9 cents went to fuel and direct fees, 18 cents went to vehicle costs and the remaining 15 cents covered corporate overhead and profit.
This in essence states that taxi business has upper limit of 15% margin. That's very low however that does not mean Uber or Lyft like businesses are fundamentally flawed. It just means Uber/Lyft has to just figure out how to survive with, say, 10% margin and they have opportunity to drive out higher margin traditional businesses. The nationwide volume can still yield great profitability (just like Amazon model). So Uber/Lyft are perfectly viable and profitable businesses, at least in theory.
Also notice that driver comp is 58%. That's massive. Self-driving will absolutely change the landscape. I would highly doubt that vehicle ownership would be as important as it is now once self-driving becomes norm. It's similar to everyone doesn't own water well any more because you have water on-demand at any time you want.
AFAICT, Horan doesn’t dispute that there is business there; he disputes that it is worth their valuation. If, say, they can make a hundred million dollars of profit a year, long term, that is a viable and profitable business, but not one that’s worth tens of billions of dollars.
And vehicle ownership? I think it is extremely hard to predict what will happen. Many people currently see their car as an extension of their home. If (possibly a big if) self-driving cars aren’t a lot more expensive, they may not want to give that up. Where else would people keep their sweets to eat during their commute, spare jacket, running shoes, tennis rackets, golf clubs, kids toys (let alone kids seats)?
There also may be extra stuff people may want to keep in _their_ self-driving car, for example a computer gaming system, a yoga mat or fitness machine (once self driving cars are safe enough, compulsory seat belts laws can go away)
Also the idea that self driving cars will end traffic is a dangerous lie to tell. Self driving cars will do nothing to reduce traffic. If anything, self driving cars will increase car trips. And in fact in order to serve people with reasonable service times there will need to be _more_ cars on the road at all times which will more than soak up any additional capacity gained from tighter traffic patterns on restricted access highways.
https://en.m.wikipedia.org/wiki/Jevons_paradox
I'm not convinced about the first idea, though. I think we shouldn't get confused with equating not covering all trips with lack of success. Even if driverless only works for commutes in one or two major cities I think that might be enough to call it a success, economical and societal.
When I first heard that it sounded crazy, but think of this scenario: you live near your friends, but go to work at different places. You all agree to meet up for a beer after work. If everyone takes the same taxi home from the pub, you've saved two car trips.
Before, driving was a job that everyone had and took up valuable cycles away from living your life. You had to consider each trip, e.g., "How much gas do I have? What's the traffic like? My car isn't working, I have to skip going out tonight."
After, no matter what you want to do, you just call a car and get in. You don't even have to think about it. It's going to liberate so many people that undoubtedly frivolous trips are going to stack, similar to how people in countries with socialized medicine tend to visit the emergency room for less and less serious medical problems.
Since Uber has come to town I catch myself calling it for things I could have easily skipped or taken the bus to or suffered on my motorcycle for (rain, etc.). I might haven taken, on average, less than one taxi per year but now take at least 15 Uber trips during that same time period. I think this sort of liberating convenience will only increase once driverless cars become popular.
Will it though? Lyft claims that self driving cars are smoother and more consistent. They claim this will end traffic. Smoother, consistent driving doesn’t take into account traffic lights, stop signs, merging, lane changing and turning, all of which require you to slow down, even for a bit, and all of which are sources of traffic congestion. And of course bottlenecks as you move from a highway to city streets, and the slowing down as you reach your final destination. Intuitively, you would need a vast, high capacity tunnel network that would allow you to avoid all those things to “end” traffic (maybe). That’s more than self driving cars alone.
Self-driving cars must follow the speed limit at all times.
How can a company program a car to break the law?
I think following the laws and driving calmly is a great benefit. Never once I have thought, "I'm so happy that Uber driver drove like a maniac weaving between lanes, and got me to my destination all of 4 minutes earlier. Yay".
Worst case scenario - the cars NEVER self-drive and always have a driver - the whole setup is exactly as economically viable as current taxis. From all reports, current taxi setups make a lot of money.
If the cars can self-drive lets say 25% of the time, then the whole setup is that much more economically viable.
Hmm... but aren't the small well-mapped (and high-density) neighborhoods where almost all the money is?
You see way more conventional cabs roaming Manhattan and the Chicago Loop than you do on a county road outside Minot, North Dakota.
Also, there's no reason they couldn't keep some human drivers for trips outside the dense core regions. The apps know where you're going before the car is even dispatched, after all.
"This guy is going from the Art Institute to the Field Museum... send him a robocar." "This guy is going from O'Hare to a farmhouse on the outskirts of Kankakee... send him a human driver."
You can’t do hundreds of millions of dollars in raises forever without some proof of progress (or maybe you can if SoftBank is willing to pour money in for a decade).
At some point, someone is going to be left holding the bag when it’s evident that the last 20% is the hardest part (but mandatory for level 5 service).
Going really extreme: in a car accident in a somewhat remote area, the driver might be able to get an injured passenger out if the passenger stuck. Now a fully autonomous car, there is only passenger. Yes, for sure this is extreme and perhaps likely a dramatic scene only, but it is one of those edge cases we can’t cover in a self-driving car.
In theory, a self-driving car could watch for hand-based turn signals, just like a human driver could, and take those into account.
In practice, what someone signals is far less interesting than what they do, and all it does is provide hints for reaction time, which 1) a computer already wins at, and 2) a computer should always leave enough time for braking/stopping/etc anyway. A human driver should never make an assumption based on someone's turn signal that would lead to a crash if they don't do what they're signaling, and neither should a computer.
> What should the machine do when there is a dog standing in the middle of a street?
Don't hit it (stopping if necessary), and listen when the human in the vehicle says "stop" (or "call animal control", or any number of other reactions). The human in the vehicle should always be in control of the high-level actions, just not the moment-to-moment ones.
> Going really extreme: in a car accident in a somewhat remote area, the driver might be able to get an injured passenger out if the passenger stuck. Now a fully autonomous car, there is only passenger.
Not sure what scenario you're getting at here. If you mean a driver and passenger in the same vehicle, the same problem exists today with an injured driver.
Edit: Instead of just downvoting and moving on, why not actually formulate a response? Is there anything more empty than disagreement without the capacity to argue your case?
I think that becomes easier, though: in the case where no human in the car has control over what the car does, it's obviously the manufacturer's fault when something goes awry. Though if doing something like rapidly and repeatedly changing the car's destination and intentionally "confusing" it leads to a crash, that might shift blame to one of the passengers, which would be unusual.
I'd agree that if we aren't seeing better fatality numbers, people won't accept it -- but there's little to suggest that humans are a difficult target to beat. Especially in the US, where car-related fatalities are so numerous and much of the research and trials are being done.
Working out accountability for fatalities is already a hugely complicated process, involving an unhealthy number of law enforcement and legal personnel. The problem of identifying responsible party for a - as you put it - failed algorithm doesn't seem intractable, and certainly isn't a show-stopper.
Haven't voted either way on this thread, but I suspect your re-hashing of regular topics already resolved is the problem.
People have this weird psychological effect around control. They will accept higher failure rates if they believe (correctly or otherwise) that they have control over a situation. Or rather, they will believe that they can beat the statistics at least for themselves, as long as they believe they have control over the situation, and achieve a lower failure rate then the driverless car. In other words, "the higher failure rate might apply to the average person, but I'm above average".
It's related to faulty risk assessment: it's much more likely that you'll get in a car and die than it is that you'll be killed in a terrorist attack, but most people don't even think of getting into a car as all that risky, and yet we start drawn-out, unwinnable wars because people are afraid of terrorism. People believe that they personally won't die in a car crash because they have control of the situation and that they're a good driver, but they have no control over when or where a terrorist attack will take place, so it's somehow a worse threat.
That's why we have courts (and insurance).
Conventional vehicles kill people all the time, sometimes through provable manufacturer negligence. But we still have them.
Then since you have proven that you could try acquiring Uber and or Lyft at a fire sale price because they couldn't come up with a solution in time and get their data and logistics.
"Let Uber drive you to work for free then pay you to use your car for the day. We'll cover the gas!"
"GM is happy to partner with Lyft to be the first self diving car! GM Good. GM Hip"
To signal to passengers that you won’t record them during the ride, give the cameras nice red LEDs when in operation, and be open about what you do.
I doubt this would even be a concern. Many Uber/Lyft drivers already record interior and exterior views for the duration of every ride. And I don't see anything wrong with that. Hell, think about:
"The rider after you reported damage to the car that you didn't report. Therefore you did it and are responsible for paying for it."
"Where's your proof? Maybe I just didn't notice the damage and it was the rider before me -- it's not my job to inspect your cars. Or maybe the rider after me damaged the car -- possibly even on purpose -- and then blamed me."
They'd be foolish not to record every single ride, from multiple angles if possible.
I am thinking like a public transit subway in New York. We cannot have anything you can't pressure wash.
However, that also brings up the problem of traffic congestion and automation (as well as comfort and ergonomics?). Can we automate the process of sanitizing a car? Will passengers accept (cold) plastic and aluminum surfaces everywhere?
I believe so too, but from a slightly different angle: I believe self-driving vehicles will literally mean the end for both Uber & Lyft (and they only accelerate their fate by investing into it).
To understand why: their inherent value is their ability to match human drivers with customers. They both do such a good job at it, that the switching costs between the 2 apps are quite low (first sign of weak competitive advantage).
Now, take away the difficult problem of finding human drivers at the right place and the right time. You now have self-driving cars that make no mistakes, have no scheduling issues, do not get tired etc. It’s way easier to fullfil demand. But, you’re not the only one working on this problem. Google, Tesla, (probably Apple too) are all actively working on the same problem. And they probably have 10X the computing/engineering/cash/data resources to solve this problem.
What are Uber/Lyft left with? A fancy app with a button (which they have already proven to easily clone). Not only that, they also have to compete with the native operating systems (and future devices) that the big corps control. They don’t even own their maps!
So yes, in 2023, you will probably be able to order a Level-5 self driving vehicle to your doorstep with a wink of your half robotic eye, but it will not be Uber or Lyft.
The way I see it, Lyft and Uber are already "big corps", and there are likely several logistic concerns that give Lyft/Uber a major head start on Google despite Google's alleged software superiority. However, I don't see Google releasing a massive self driving fleet in San Francisco and simultaneously capturing significant market share overnight.
Tesla on the other hand has hundreds of thousands of cars on the road and is rapidly on their way to 1M cars. Data won’t be as raw and abundant, but that doesn’t mean it isn’t potentially very useful none the less.
Not sure how close this puts them in the race.
Simply putting more SDCs on the road does not guarantee they will focus on the data they need.
Agree completely
It’s not just cash. Who has been building cars with smart sensors for more than 10 years? Who has self-driving cars on the road already? Who has factories to scale this up to a level that’s required to build an enormous fleet to replace all existing cars? It doesn’t even come close.
It may (read: will and has) take longer than they wished, but they aren't in anywhere near the danger of going under as they were with, say, the original Roadster, so this is just a matter of time, in my opinion.
The danger comes in the fact that electric cars are ultimately fairly simple to make (although Tesla does a particularly good job at it), and if their competitors just got their act together, Tesla would have a tough time. But in my opinion, as long as they're the only ones with a really good charging network (the others are both much smaller and much slower... and the few stations that aren't slow require an upgraded car), they have a significant head start on their competitors.
Production bottleneck will eventually clear. But having demand exceed supply is really, REALLY good for Tesla's business as it gives them access to interest-free capital (from reservations) and the reservations also show the existence of sufficient demand to justify financing continued expansion. So, Tesla keeps coming out with new products even before the previous bottlenecks have been cleared in order to maintain the state of having demand greatly exceed supply.
And these new products usually share a technology core, i.e. battery modules and electronics, thus allowing them to traverse as far down the learning curve as is possible, which is a multiplicative effect that improves every other product. Using larger batteries than you might think strictly necessary only amplifies this effect while giving them a quality and performance advantage over their competitors.
Tesla will also have an advantage in their charging network of having cheaper electricity with faster charging than anyone else. The Semi product offers 7 cents per kWh, which is lower than residential rates just about anywhere AND is ultra fast-charging. So assuming they can meet that rate, Tesla will have a large advantage in logistics costs even compared to other electric self-driving providers.
I think GP was being a bit hyperbolic, but I think (but don’t know for sure) Tesla might have 10x the telemetry data. Uber could pass them up though very rapidly if they made some changes to how drivers use phones and what data they collect (collecting video and accelerometer / gyro data for example).
Side by side as of a few minutes ago.
> Uber owns their maps.
That seems rather definitive compared to:
> And where did I claim it does not currently use Google Maps?
Don’t say things in matter of fact form and then push the blame on the questioner when evidence is shown that your “fact” seems anything but.
Yes but that app is branded Uber/Lyft, and while building an app with a button is easy, when was the last time you tried searching Bing or DuckDuckGo because Google Search was returning nonsense?
There's also question of mindshare; getting people to install an app and create an account is difficult enough, never mind B2B business contracts, such as uberpool being a recognized WageWorks commuter benefit, that are business-y to replicate.
There have been many solid competitors (both hardware and software) to the iPhone, but there's still only one Apple.
[1]: https://techcrunch.com/2017/12/06/lyfts-self-driving-pilot-w...
In more seriousness, I'm curious how well this is supposed to play with the other partnerships that Lyft already has? i.e. all together, is this suppose to represent one coherent effort to build L5 autonomous, or multiple parallel ones?
When self driving cars become a reality, I don’t see the technology or the data becoming a real sustainable competitive advantage for any one player. Once the last technical hurdles have been overcome and society has adapted in whatever way it needs to, it just won’t be that difficult for every auto manufacturer to figure out how to get in on the game.
Fragmentation across local markets also has to come into play at some point. It’s only a matter of time before cities and towns get their acts together and figure out how to extract revenue to at the very least cover the real but indirect costs they are incurring as a result of ridesharing. They could pretty easily make it next to impossible for their non preferred vendor partners to operate in a city.
No source citied on this or any of the other assertions made on the landing page.
Siri's potential problem field is huge, driving is a well defined, limited problem field.
Many researchers I've talked to believe the initial self-driving car deployments to the public will be have to be very constrained: fixed routes, possibly isolated self-driving car lanes, only allowed in clear/daytime conditions.
Essentially train carriages on tarmac :)
You should be asking yourself what will happen to the average lifespan of those who won't die prematurely in a car accident.
But that doesn't stop me from wanting to know what the US will look like in a future with self-driving cars. Will we, for example, find such a severe shortage of donor hearts that certain unlucky people live less long? Will we prioritize research that could improve treatment diseases which, untreated, lead to organ transplants? Or will we consider changing laws to allow donors to be compensated for organ transplants? (which have all kinds of icky side effects but could help avert a crisis if there were one).
I genuinely don't know what the effects of fewer road deaths will be on transplant wait times and wonder how this externality (if there is one) will be handled.