35th edition of the 6th year of SmartDrivingCars
Saturday, August 18, 2018
Uber’s Losses Mount at Self-Driving Car Unit
A. Efrati, Aug 15, “Uber has been spending between $125 million and $200 million a quarter on its self-driving car unit over the past 18 months, The Information has learned, equivalent to between 15% and 30% of the company’s quarterly losses. The previously undisclosed spending highlights the financial burden that self-driving car development has imposed on Uber and why CEO Dara Khosrowshahi is under pressure to decide what to do about it.
Some investors have told Uber officials that it may be wise to divest the self-driving car unit, said a person familiar with the issue. Uber has invested least $2 billion in the unit over the past three years. Yet the company hasn’t yet come up with a clear path to commercializing the technology it has developed.
The group’s quarterly cash burn of $1 million to $2 million per day has been particularly high during quarters when Uber paid for expensive hardware like cars and sensors that are attached to the cars, said a person with knowledge of the data. The company has tried to reduce some expenses by withdrawing operations from Arizona and cutting a development effort for self-driving trucks. Broadly, though, there’s no sign that the unit’s cash needs will meaningfully come down. An Uber spokeswoman did not have a comment….
The argument against selling the autonomous unit would be that Uber needs to have a way to develop self-driving cars if other companies won’t partner with it. Long term, self-driving cars could help Uber’s ride-hailing network reduce costs from not having to employ drivers. If Uber doesn’t develop the cars itself, it will need to figure out how to get other car developers to agree to use their vehicles to pick up Uber’s passengers….” Read more Hmmmm…. They are really between a rock and a hard place. They can’t really grow without. If someone else is successful at making it work, they’ll operate it themselves rather than license it (Netflix didn’t license its service to Blockbuster). With that competition, Uber’s valuation goes to 10x earnings which is an ugly IPO . The only upside rests in Driverless, so they can’t get out. Alain
Smart Driving Cars Podcast Episode 52
F. Fishkin, Aug 18, “Uber’s future is linked to it’s ability to deploy self driving vehicles. That’s what Princeton University’s Alain Kornhauser says in Episode 52 of the Smart Driving Cars Podcast with co-host Fred Fishkin. Also…the latest from Ford, Tesla, Elon Musk, Kroger and more. Tune in and subscribe!” Hmmmm…. Now you can just say “Alexa, play the Smart Driving Cars podcast!” . Ditto with Siri, and GooglePlay. Alain
Real information every week. Lively discussions with the people who are shaping the future of SmartDrivingCars. Want to become a sustaining sponsor and help us grow the SmartDrivingCars newsletter and podcast? Contact Alain Kornhauser at firstname.lastname@example.org! Alain
Ford takes an anti-Uber approach to self-driving cars
S. Dent, Aug 16, “As Uber and Tesla have learned, it doesn’t take a lot of bad press to shatter public trust on autonomous vehicles. Ford has taken that lesson to heart, saying it would rather instill confidence in self-driving cars than be first to market. In a letter to the US Department of Transportation (DoT) and 44-page report called “A Matter of Trust,” the automaker detailed how it plans to safely test its self-driving vehicles on public roads.
Why the emphasis on trust? The death of a pedestrian involving a Uber self-driving vehicle with a safety driver in particular seemed to affect the public’s opinion of self driving cars. Prior to that, 44 percent of US adults said they’d be okay riding in autonomous cars. However, a more recent Pew survey showed that three-quarters of participants would rather not be driven by robots, and half said they’d never buy a self-driving car.
Considering the rush to develop the tech, that’s pretty bad news. Automakers and tech companies like Mercedes, Toyota, Waymo and others have heavily invested in the tech, and begun testing vehicles on public roads. Toyota and NVIDIA, however, temporarily halted testing following the Uber accident.
Ford provided a fair amount of detail about its efforts, showing how it would train safety drivers and handle system malfunctions, hacking and accidents. All of its self-driving cars will have black box-like “Event Data Recorders,” much as airliners do. “It’s about a holistic strategy to reduce the risk of failures and protect people in case something does go wrong,” Ford said…” Read more Hmmmm…. Nice summary. See next 2 postings for the source documents and my comments. Alain
A Matter of Trust: Ford’s Approach to Developing Self-Driving Vehicles
S. Marakby, Aug 15, “Earlier this week, I shared a report outlining our approach to self-driving vehicle development with U.S. Department of Transportation Secretary Elaine Chao. Our full report, “A Matter of Trust,” is available for you to read and you can view the letter I sent to Secretary Chao below.
As Ford pursues its goal of developing a self-driving vehicle for deployment at scale in 2021, we believe that ensuring people trust the safety, reliability and the experience that our vehicles will enable is paramount. We’ve spent more than a century earning consumer trust by designing and building safe, quality vehicles, and we’re committed to preserving that trust in a future of self-driving vehicles.
In our report, you will learn how we always prioritize safety at Ford; how we are working closely with industry and government partners; and how we are applying self-driving technology to solve the challenges our cities face. Our goal is to give you a sense of the new services that will be available in the near future and how they could impact our lives…” Read more Hmmmm…. Read: A Matter of Trust: “…This growth is also causing a shift from individual vehicle ownership to the use of shared mobility options such as ride-hailing services.” …. Nice…! “Most of our infrastructure was built to meet the needs of individually used vehicles. However, most of those vehicles sit idle about 95% of the time. As a result of this, as much as 30% of the real estate in city centers is devoted to parking. “ …. Not so nice. Sure, reducing parking is nice, but the real value of Driverless is Shared riding. That really improves utilization; reduces, on a proportional basis, energy, pollution and affordability; and also congestion and, finally, parking. With such enormous societal benefits “… to help make people’s lives better….” , promoting and operating this new technology to achieve ride sharing, if only during peak hours in peak directions and only in twos and threes, would seem to be perfect for such an “…enabling the future…” vision statement. Real “ride sharing” ( normal Uber/Lyft is NOT ride-sharing. It is “ride-chauffeuring”… great as long as energy, environment, congestion and your pocketbook can afford it. ) is where the real societal benefit kicks in.
To be fair, the report does emphasize that Safety is an absolute necessary condition and “…Can I buy a Ford self-driving vehicle in 2021?
No, our self-driving vehicles won’t initially be sold to customers in the way that cars are today. You’ll be able to experience these vehicles through multiple means, commercial fleets in mobility services such as ride-hailing and goods delivery….” Excellent!!
The report goes on to concisely describe very well the many operational characteristics. Its major shortcoming is the lack of appreciation of ride-sharing.(which is in line with the rest of the OEMs). Its major, positive in my view, differentiator is their exclusive focus on fleet operations and not on consumer sales. Does this mean that they would sell/lease Driverless vehicles to Uber, as they sell specialty vehicles to police and taxi, or choose instead to capture the profitability upside of the technology by deploying and managing the vehicles themselves (since they don’t need to hire “policePerson” or “taxiPersons”) is not addressed. it would all depend on how much of a premium Uber is will to pay to lease these specialty cars. That premium must be equivalent to the profitability of operating the cars themselves which Uber can’t afford to pay. So Ford operates them and Uber needs to make teir own. Alain
A Waymo engineer told us why a virtual-world simulation is crucial to the future of self-driving cars
M. DeBord, Aug 16, “Waymo has racked up over eight million self-driving miles since the former Google Car project kicked off almost a decade ago. But that impressive real-world number is nothing compared to the mileage amassed in virtual reality.
“We do eight million miles every day in simulations,” said James Stout, a senior staff software engineer at Waymo….” Read more Hmmmm…. Impressive; however, it is not the sum of the miles, but the “quality” or challenge of the miles. One learns a lot about experiencing, in reality or virtual reality” “challenging miles” (corner cases that may involve different variations on what is other wise just a few hundred feet) and learn absolutely nothing from millions of repetitive trivial miles where nothing challenging is being encountered. Yes, virtual reality is a very efficient, effective and robust environment for stress testing and improving automated driving systems. Alain
Why Tesla Dropped Nvidia’s AI Platform For Self-Driving Cars And Built Its Own
J. Su, Aug 15, “… similar strategy that Apple adopted when it decided to design its own mobile chip.
“We’ve been like semi-stealth mode basically for the last two to three years on this,” admitted Musk during the company’s Q2 earnings call. The electric carmaker openly hired recognized experts in the field including legendary chip architect Jim Keller, formerly at AMD and Apple and who’s now heading Intel’s AI chip development.
We believe that the main reasons for Musk’s decision are a combination of performance and control, with the price of the custom chip being the icing on the cake as it will end up costing less than the current Nvidia chips…. ” Read more Hmmmm….Seems somewhat logical if one doesn’t have other “cats to heard”. Given what is on Tesla’s plate, this seems to be something it can/should be letting someone else manage. Alain
Short seller sues Tesla, Elon Musk, claiming buyout eets were fraudulent
T. Lee, Aug 11, “Three days after Elon Musk tweeted that he had “funding secured” to take Tesla private, we still don’t know who, if anyone, has agreed to provide the billions of dollars needed to buy out the company. But plaintiffs’ lawyers aren’t waiting any longer for Musk to provide more details.
On Friday, stock trader Kalman Isaacs filed a class-action lawsuit arguing that Musk’s Tuesday tweets constituted securities fraud. The lawsuit, first reported by Reuters, appears to be the first one claiming that Musk’s Tuesday tweets violated federal securities law.
At the time of Musk’s tweet, Isaacs was short at least 3,000 shares of Tesla stock. That means that, at some point prior to Musk’s tweet, Isaacs had borrowed Tesla shares and sold them, betting that the price would fall and he would eventually be able to buy them back at a discount, pocketing the difference.
The flipside, however, was that Isaacs lost money if Tesla’s stock rose—$3,000 for every dollar increase in the stock price. If Musk really did close a $420-a-share buyout deal, Isaacs would have needed to come up with around $420 for each of those 3,000 shares. So on Wednesday, Isaacs panicked and closed out his short position early by buying Tesla shares at the then-current price of $376….” Read more Hmmmm…. Poor Elon. He shouldn’t “Tweet out loud”. Alain
Artists, Accessibility and Autonomous Vehicles – Part 2 #AVS18
K. Pyle, Aug 13, “A long-time proponent of autonomous taxis, Princeton Professor Dr Alain Kornhauser discusses a potential downside for this future mobility as a service offering. Speaking at the AVS18 Symposium, Kornhauser warns that operators will need to account for how people treat property they don’t own. In the above video, he provides many examples of graffiti-covered public transportation systems. The flip-side of graffiti is art and he suggests that shared autonomous vehicles have the potential to be a sort of mobile canvas that could be a community asset, beyond better mobility.
To get to a world of mobile art, instead of rolling graffiti, Kornhauser suggests that RESPECT will be key….(Respect In Peace Aretha and thank you!!!)” Read more Hmmmm….also see video. Alain
Designing for Autonomous Accessibility
K. Pyle, Aug 17, ““Our focus is on making sure that transportation of all kinds is available for people with disabilities and older adults and their caregivers,” explains Carol Wright Kenderdine, Co-Director of the National Aging and Disability Transportation Center (NADTC). A joint venture of the Easterseals and the U.S. D.O.T.’s Federal Transit Administration, the NADTC works with transit agencies to improve transportation accessibility, whether “walking, biking, driving, and transitioning from driving, as well as fixed-route public transit, dial-a-ride services, volunteer transportation programs, taxis and other shared ride options.“…She stresses that autonomous vehicle manufacturers and service providers reach out early to groups that the NADTC represent to understand the needs and design accordingly…” Read more Hmmmm…. also see video. A substantial difference about autonomousTaxis is that they are individually managed and dispatched to serve a unique demand. The right vehicle can be assigned to serve the right demand. Consequently, I hope that it will be OK with everyone if specially equipped vehicles can be used to provide mobility to those who can benefit and need the special equipment and not that each vehicle must have all of the specialty equipment so that it can be used by anybody. All that should be required is that the level-of-service offered (basically wait time) to any customer is independent of the special equipment. This can be achieved if a sufficient quantity of each variety is available for dispatch. Unfortunately, that may not have been Carol’s focus. Alain“
WAYMO OR BAIDU & CO. – WHO WILL WIN THE ROBOT CAR RACE?
J. Becker, Aug 16, “..The question remains who will win the race in the end. Baidu will dominate all others in China, that much is foreseeable. In the rest of the world, Google (Waymo) wants to be the big winner – after all, it was the same with Android. The U.S. is still the world leader in artificial intelligence. And for some time now, Sundar Pichai, head of Google’s parent company Alphabet, has been preaching “AI first”…” Read more Hmmmm….Unfortunately, the question is not answered, except everyone knows the answer… As is stated Baidu will win in China but they won’t have the bandwidth to export because they can’t afford to screw up screw up in China. … Waymo will win in the US and will be unable to crack the Chinese market. Waymo will also find substantial pushback in Europe and India from local entrenched interests. Europe will lag behind for years thwarted by the conventional auto industry protecting the status quo. Just a thought. Alain
WHICH CAR MAKERS WILL SURVIVE THE NEW MOBILITY?
Alex Roy, May 23, ““In the future,” said Soichiro Honda, “there will be just a half dozen car companies. And Morgan.”
That’s an incredible statement, from one of the most brilliant minds in human history, because 1) it assumes no level of innovation or profit can save 10 or more of the largest manufacturers in the world from bankruptcy and/or consolidation, and 2) Morgan, the oldest privately held car maker in the world – notorious for its reliability and production delays – will survive despite a total lack of innovation. What possible catalyst could precipitate this schizophrenic apocalypse?
Doom, thy name is mobility, and those who fail to understand it are toast. …The New Mobility is therefore the elimination of inefficiencies and friction on the transportation continuum, both within and in between verticals….” Read more Hmmmm…. Very interesting and thought provoking. Alain
TLC flooded with applications before ride-share cap kicks in
D. Furfaro, Aug 10, “Uber is causing traffic jams again — at the TLC office!
The ride-sharing giant offered $500 rewards for new drivers to get their cars licensed by the Taxi and Limousine Commission before the city slammed the brakes on its vehicle numbers — leading to a flood of last-minute applications. The TLC says it received a whopping 3,152 for-hire vehicle applications between Aug. 1 and Aug. 9 alone — 560 more than it did for the entire month of August last year. …” Read more Hmmmm….Thank you Uber for providing a $500. subsidy to add another vehicle that delivers un-subsidized “paratransit service” to New York City. )How much does it cost the public to have the MTA add a paratransit vehicle and deliver rides???… Why Paratransit Doesn’t Have to Be So Expensive: …”New York’s Metropolitan Transportation Authority (MTA) spends more than $70 per paratransit trip and has high per-capita paratransit usage. As a result, the MTA subsidized its Access-A-Ride service last year (2015) to the tune of $256 million…”) Alain
Iowa among states barring automated truck ‘platoons’
M. Schmidt, July 24, “Iowa is one of 34 states that does not allow for automated vehicle platooning — a budding technology that transportation officials say provides major benefits to the movement of freight. A new report released this month by Competitive Enterprise Institute, a Washington-based public policy organization, indicates that following-too-closely laws are the most common legal barriers to vehicle platoons.
In a platoon, two trucks — with onboard automated technology — can link up to share a signal. Once connected, the lead truck sets the speed, while the rear truck follows as close as 40 to 50 feet behind. With the shared signal, the rear truck’s automated braking system can sense the forward vehicle’s brakes to slow down or stop accordingly. …” Read more Hmmmm….This application is not Automation’s “lowest hanging fruit”. Safety is the “lowest hanging fruit” and is THE necessary condition for any other application of automation (taking hands off wheel, feet off pedals, platooning, taking drivers out of trucks…). So the motor carrier industry should focus on using automation to improve SAFETY. With Safety achieved, lead trucks don’t crash or run off the road anymore (or not nearly as often). Once we have Safe-driving Trucks, then we can bring home platooning. Although, if anyone is really serious about platooning, the best first application is to so equip the bus fleet using the XBL (Exclusive Bus Lane) and the Lincoln Tunnel leading to the PA Bus terminal in NYC. There the benefit is the ability to move 17,500 more people per hour into midtown ( a 50% increase) without having to build more roadway infrastructure (which would cost ~$10B v the automation technology costing ~$10k/bus * 3,000 buses = $30M) . Now that’s a tangible benefit!! Originally proposed by A. Kornhauser & L. Pignataro, Operational Test of AHS Applied to a High-Capacity Transit Corridor, The Lincoln Tunnel XBL, October 1995! Alain
To Get Ready for Robot Driving, Some Want to Reprogram Pedestrians
J. Kahn, Aug 16, “In a line reminiscent of Steve Jobs’s famous defense of the iPhone 4’s flawed antennae—“Don’t hold it like that”—these technologists say the problem isn’t that self-driving cars don’t work, it’s that people act unpredictably. “What we tell people is, ‘Please be lawful and please be considerate,’” says Andrew Ng, a well-known machine learning researcher who runs a venture fund that invests in AI-enabled companies, including self-driving startup Drive.AI. In other words: no jaywalking…
With these timelines slipping, driverless proponents like Ng say there’s one surefire shortcut to getting self-driving cars on the streets sooner: persuade pedestrians to behave less erratically. If they use crosswalks, where there are contextual clues—pavement markings and stop lights—the software is more likely to identify them.
But to others the very fact that Ng is suggesting such a thing is a sign that today’s technology simply can’t deliver self-driving cars as originally envisioned. …”
Read more Hmmmm….What makes this not Half-baked or C’mon Man is because there is real substance in this topic… (and bicyclists and powered scooters and …) will need to behave and obey some rules at least as well as they obey them today. Today, pedestrians and bicyclists are not permitted on freeways because human drivers would readily kill them there. Fences are placed on top of concrete medians on many sections of US 1 so as to discourage pedestrian from trying to cross. Similarly, fences are placed on the edge of pedestrian overpasses to discourage kids from dropping “bowling balls” on top of the passing cars below. It is trivial for misbehavior to bring today’s conventional roadway system to it knees. Certainly pedestrians today behave differently today then in the 19th century prior to the introduction of the conventional automobile. Also, the Elaine Herzberg crash would likely not have been averted had the car been a conventional car operated by an alert driver. The lighting conditions and vehicle speed were such that even an alert driver would not have seen her in time. In fact, that was the initial finding of the police investigation. So, automation’s superior sensors and tenacious operation will enable it to be more tolerant of some misbehaviors; however, certainly not all misbehaviors. There will need to be clear “rules of the driverless road” that the driverless vehicles will need to obey as well as everyone else, otherwise, people are going to get hurt and some will die. That number will be less than today’s, if the level of misbehavior is constant, but could easily be more, if misbehavior explodes. If you jump in front of a driverless car, it will kill you and it will be your fault, not the driverless car’s fault. That’s why misbehavior will need to be addressed very seriously.
And, of course the AI needs to be able to understand body language and gestures and anticipate some misbehavior much as we each do today when we drive. (When I drive down “Nassau Street” I try to determine if the traffic side of a parked car that I’m approaching is occupied. If so, I tend to give a little more space in case that person decides to fling open a car door. Why…, because, a long time ago, someone did that to me and I hit it. So now I try to give myself a little more room in case someone does that to me again. Such behaviors are/will need to be baked into the AI of these vehicles. Alain
Kroger Tests Driverless Cars for Grocery Deliveries
AP, Aug 16, “At a time when big-box retailers are trying to offer the same conveniences as their online competitors, the grocery chain Kroger is testing the use of driverless cars to deliver groceries in a Phoenix suburb. The project from Kroger, the biggest supermarket chain in the United States, is part of an effort to fend off a rising threat from Amazon and Walmart, which have been pushing their online grocery offerings.
Kroger’s pilot program started on Thursday with a robotic vehicle parked outside one of its Fry’s supermarkets in Scottsdale. A store clerk loaded grocery bags into the back seat of a car with two men in the front seats, one with a laptop. Both were there to monitor the car’s performance…” Read more Hmmmm….Probably should be in Half-baked… When you have “two men” in the front seat, it can’t be Driverless. I understand that one want’s to test the concept, but one doesn’t need the driverless features to test the concept of home delivery of groceries. Seems like a very expensive way to test Driverless. What am I missing??
Uber Drivers Take Riders the Long Way—at Uber’s Expense
G. Bensinger, Aug 13, “When ferrying passengers to and from Phoenix’s airport, Uber driver Michelle Blandy had a decision to make: Go the short way or take the longer route.
“If they were from out of town, I would take advantage of it by going the longer route,” said Ms. Blandy, who drove for Uber for about two years before moving to Harrisburg, Pa., earlier this year. “It’s the only way I could get what I was owed.”
In a modern twist on an age-old practice from the yellow cab industry, many drivers like Ms. Blandy are employing a practice known as longhauling—taking an unnecessarily longer route to a destination in order to drive up a fare. But unlike with taxis where passengers “get taken for a ride,” it’s ride-hailing companies like Uber Technologies Inc. that are responsible for covering the bigger bill….” Read more Hmmmm….I had a Lyft driver try to do this to me London to Gatwick. Good thing I had CoPilot 🙂 This issue won’t go away with autonomousTaxis. The fleet operator has similar incentives. This will require regulation. it should be something that the California PUC should include in their regulations in addition to allowing/encouraging shared-rides. Alain
Interested in working in Toronto? Have a good background and interest in working on safety and security for autonomous driving vehicles and fleets? Contact Dr. Fengmin Gong, DiDi Labs
Half-baked stuff that probably doesn’t deserve your time
C’mon Man! (These folks didn’t get/read the memo)
Calendar of Upcoming Events:
3rd Annual Princeton SmartDrivingCar Summit
evening May 14 through May 16, 2019
Save the Date; Reserve your Sponsorship
Catalog of Videos of Presentations @ 2nd Annual Princeton SmartDrivingCar Summit
Photos from 2nd Annual Princeton SmartDrivingCar Summit
Program & Links to slides from 2nd Annual Princeton SmartDrivingCar Summit
On the More Technical Side