Uber fires Anthony Levandowski amid self-driving tech lawsuit with Waymo

Uber on Tuesday announced the immediate firing of Anthony Levandowski, one of the leaders of its self-driving car efforts, amid a lawsuit accusing it of using data allegedly stolen from Alphabet’s Waymo unit.

Levandowski has remained silent, refusing to cooperate with either Uber or a federal judge, choosing instead to invoke his Fifth Amendment rights.

Apple’s own self-driving efforts have been largely kept in the dark. The company has started public testing using modified Lexus RX450h SUVs, but it remains unknown if the company intends to partner with an existing automaker or design its own vehicle. That decision could happen by the end of 2017.

appleinsider – Uber fires Anthony Levandowski amid self-driving tech lawsuit with Waymo

[HT mikeyp]

Self-Driving Uber Apparently Runs Red Light In San Francisco

“This incident was due to human error,” the statement read. “This is why we believe so much in making the roads safer by building self-driving Ubers. This vehicle was not part of the pilot and was not carrying customers. The driver involved has been suspended while we continue to investigate.”

State regulators said Wednesday afternoon that Uber must stop the self-driving car service until it receives a permit from the state.

via CBS SFBayArea – Self-Driving Uber Apparently Runs Red Light In San Francisco

HT kanaad

Report says Uber surge pricing has a twist: some drivers flee – SFGate

When his Uber app notifies him about a surge, a price increase in an area where rider demand is high, Sollars knows just what to do: He drives his black Ford Escape somewhere else.
“The seasoned drivers don’t pay any attention to surge,” he said. “By the time you get to that part of the city, the surge is over. Often, even when I’m sitting dead center in the middle of a surge area, I don’t get a ride request. Then, as soon as the surge is off — bam! — here comes a ride.”

Statement:

Still, by quashing demand, surge ensures that passengers who are willing to pay more can get a ride quickly, achieving its basic goal. “One side is working; it definitely impacts demand, but the impact on supply is minimal,” Wilson said.

Counter:

“They missed a lot of what surge does with respect to supply,” said Keith Chen, a UCLA associate economics professor currently on a two-year leave to work with Uber as its head of economic research, designing the third iteration of its surge system. During busy times, the study’s method would miss cars that “got gobbled up” by riders and instantly replaced by other cars, he said.

Idea: perhaps Uber should provide an estimated duration for the surge to its drivers.

via Report says Uber surge pricing has a twist: some drivers flee – SFGate.

Paper: Peeking Beneath the Hood of Uber, IMC 2015

Efficiencies and Regulatory Shortcuts: How Should We Regulate Companies like Airbnb and Uber? – Working Paper – Harvard Business School

Enjoyed reading a comprehensive working paper from the Harvard Business School on the upsides, downsides, and potential downsides of two-sided markets (aka platforms); implications and challenges for the legal system with respect to regulating these new types of services. The paper suggests ending “protectionist” regulation and discusses negative externalities (many of which have initial evidence but require further study), asymmetry in information (as compared to incumbent services), cognitive biases such as racial profiling based on profile picture, not being incentivized to provide “universal service”. The paper provides a nice overview of both driver and consumer perspectives, concerns, and challenges.

Abstract:

New software platforms use modern information technology, including full-featured web sites and mobile apps, to allow service providers and consumers to transact with relative ease and increased trust. These platforms provide notable benefits including reducing transaction costs, improving allocation of resources, and information and pricing efficiencies. Yet they also raise questions of regulation, including how regulation should adapt to new services and capabilities, and how to correct market failures that may arise. We explore these challenges and suggest an updated regulatory framework that is sufficiently flexible to allow software platforms to operate and deliver their benefits, while ensuring that service providers, users and third parties are adequately protected from harms that may arise.

It appears that the potential and role of government in regulating these services is to provide sustainability of these services (whether through private or public, one or many players); this means long-term and consistent reliability and wide-spread and fair access. The role of government is to look beyond the capitalistic systems upon which corporations operate and look beyond the short-sighted-ness of individual citizens, as a way to protect it’s people.

Regulation can usefully set minimum standards to protect consumers who fail to recognize potential problems and to protect against problems prior consumers could not notice… Many long standing transportation requirements address aspects of safety that customers would struggle to access even after a ride–for example, requiring vehicle inspection with heightened frequency or rigor.

What is role of statistics and control theory in these questions of providing assurances to the consumer?

via Efficiencies and Regulatory Shortcuts: How Should We Regulate Companies like Airbnb and Uber? – Working Paper – Harvard Business School.

Citation: Edelman, Benjamin G., and Damien Geradin. “Efficiencies and Regulatory Shortcuts: How Should We Regulate Companies like Airbnb and Uber?” Harvard Business School Working Paper, No. 16-026, September 2015.

Getting Over Uber — Backchannel — Medium

Susan Crawford (Professor of Law) promotes improving taxis in urban areas, as a way to provide reliable and cost-effective mobility to people in general. Uber aims also to provide reliability, but the definitions are fundamentally different.

This following trends makes sense, given the incentives (and purpose) of corporations to produce profit:

Uber drivers have a tough time making a living; they’re responsible for their own cars, fuel, benefits, maintenance, tolls, and certain insurance as well as the kickback to Uber that takes a substantial slice out of every fare they pick up. They may or may not know where they’re going, and they may or may not be driving cars that are safe. Uber consistently squeezes its drivers as tightly as it possibly can; new drivers are paying an even higher cut to Uber than the first generation did.

These trends will be particularly true for the lower-income drivers, who rely on driving as their primary income.

It would be interesting and insightful to do a historical analysis on the AT&T monopoly in telecommunications and draw parallels to this pending monopoly in transportation networks.

So even when private companies provide basic transport and communications services — in America, that’s often how we do it — they do this subject to extensive public obligations. That’s where the whole idea of “common carriage” came from — transport and communications networks operated by private companies that provided a high level of uniform service at uniform rates under uniform promises of safety and reliability.

I’m not sure that taxis are the answer, however, though the model of Mobility-on-Demand (MoD) systems is extremely compelling in urban areas, not necessarily operated by a single player (or even a single private player). As some commenters point out, taxi drivers also face many of the same issues as Uber drivers. Taxis are an old and slow-moving system, not very suitable to “compete” in today’s fast-paced world. Governments and public agencies must be willing to experiment, break their own rules, and move fast, if they wish to protect its people.

One last curiosity: bus drivers in some American cities are paid well and have nice benefits. They welcome riders with a smile and some centralized system optimizes their shifts so they have variety in their weekly routes. Perhaps there is something to be learned from public transit systems when considering MoD systems.

via Getting Over Uber — Backchannel — Medium.

A very nice response by Tim O’Reilly addresses many additional problems and differences between taxi and Uber, most notably access to a private vehicle. (Taxis are rented; Uber drivers are required to have their own vehicle.) He lays out several actionable items for the government:

1. Removing outdated taxi regulations that make it difficult for taxis to compete with Uber and Lyft, even given comparable technology. For one example, consider Washington D.C., served by taxis with geographic restrictions. A taxi driver can pick up a passenger in Montgomery MD and bring her into DC, but must then drive back to Maryland to pick up another passenger, since he is prohibited from picking up in the city. A driver from Virginia can only pick up Virginia-bound passengers at Washington National Airport. And so on.

2. Working closely with Uber and Lyft to understand how well the city is being served. There is some evidence that Uber and Lyft are improving availability in previously under-served neighborhoods. Cities should be working to build on and verify these studies.

3. Understanding whether the reputation systems (and other self-regulatory regimes) of Uber and Lyft are producing results at least as good as the older regulatory regimes under which taxis operate. The passenger experience suggests that they are doing considerably better than the older regulatory regimes, but cities should actively be pursuing data to confirm or disprove this anecdotal evidence, and introducing regulation only when systematic problems have been uncovered.

4. Working with Uber and Lyft to understand the tradeoffs between lower fares for passengers and driver income. There is a risk that in pursuit of low prices for consumers, these companies could end up exploiting workers. Government does have a role in making sure that companies produce great experiences not just for their customers but also for the workers delivering their services. But guess what: government has abdicated that responsibility in low wage industries like retail and fast food, where workers are paid so little that they must supplement their wages with public assistance. (Recent estimates put the taxpayer subsidy to these industries at $153 billion/year.) I’d much rather see government focus on areas like this where there is a clear and present problem rather than in new industries like on-demand transportation where the market has not yet settled on the right balance between value to customers and value to workers. The fact that Uber and Lyft are competing so hard to attract drivers suggests to me that the market still has a lot to say about that balance.

5. Improving crime reporting so that there is a consistent basis for evaluating the relative safety of taxicabs versus Uber and Lyft. While there are many anecdotal accounts of bad Uber experiences, there are also anecdotal experiences of bad taxi experiences, but crime statistics are not reported in a way that allows cities to understand if new safeguards are needed.

via Getting Over Uber — Backchannel — Medium.

A New Partnership With Europe | Uber Blog

This year, in close partnership with European cities, we can take 400,000 cars off the road, expand UberPOOL and reduce emissions, all while creating 50,000 new jobs across the continent.

As with our recent initiative with the city of Boston, Uber can share smart data with partner cities to help them manage growth, reduce congestion and greenhouse gas emissions and expand public transportation. We’re confident this combination of data sharing, job creation and reduced use of personal automobiles represents important partnership opportunities for Europe’s cities as well.

via A New Partnership With Europe | Uber Blog.

Uber is moving fast and breaking stuff, woo!

Uber’s legal battles and robocars | Brad Ideas

the Online Ride drivers make the very common mistake of acting as though the depreciation on their car costs them nothing because they have already bought the car. In talking to drivers, I have found that most do not account for the true cost of operating their car (around 50 cents/mile or about $20/hour) and rather mostly look at the cost of gasoline. They figure their income by just taking what they are paid and deducting very immediate costs like fuel. They are making an error, and this allows the companies to get drivers for less than the real cost. Robocars, of course, won’t have this issue.

Attacks on the online ride industry will continue. The official reason for slowing it down will be a reasonable sounding one such as safety

As a postscript, I should note that Uber has also gotten into deserved trouble because of very bad privacy practices and abuses, and disturbing attitudes by management about the press, their opponents and even women. These issues are unrelated to the real question, and Uber deserves trouble for them. Though it’s enemies will seize on its real mistakes in other fields to fight their battle.

via Uber's legal battles and robocars | Brad Ideas.

The battle against inertia.

The War Between Uber and Lyft Will Be Won With an Algorithm – CityLab

Lyft’s new VP of data science, Chris Pouliot, previously led the data analysis team at Netflix, which is known for utilizing the heck out of its user data. He’s got big plans for Lyft, too. "We can use data to provide more accurate [estimated times of arrival] when a passenger requests a ride, to set high expectations and provide a better user experience," Pouliot told VentureBeat in December. "The success or failure of the business is highly correlated to how the company uses data." He also hopes to make an algorithm that predicts the likelihood that someone nearby will need a car in the next three minutes. That way, each individual is sent a ride that’s convenient for them as well as the entire user base. The drivers-as-friends shtick might get a software boost, too: Pouliot imagines that a Facebook connection could tell passengers about interests they have in common with their driver.

Uber has a current job listing for a "leading data scientist," so it could be that they’re still trying to get their own Pouliot-level analytics whiz.

via The War Between Uber and Lyft Will Be Won With an Algorithm – CityLab.