Uber’s C.E.O. Choice Faces a Question of Ambitions
nytimes.com
nytimes.com
He's tasked with righting a ship that's leaking everywhere–a board embroiled in litigation, a former CEO still in a position of power, and resenting his ouster. Employees who, by now, are self-selected to consider any moves to heal the toxic culture as assaults on themselves.
Meanwhile, they've spent $1200 of investors' money on every user they currently have (including the expensive failed adventures in China and Russia). It isn't quite clear if they're profitable anywhere, or if they're continuing to be a program to distribute wealth from the bottom (drivers) and top (investors) to the middle class rivalled only by the New Deal.
What's still ominous is the complete lack of a moat around their business. Every driver and every customer is a free app install away from giving any competitor a shot. Their autonomous car project is in deep trouble–even if the Waymo lawsuit doesn't kill it, it now seems certain that a lot of assumption were optimistic–exemplified by Apple's demotion of its competing project, and the fact that the far superior technical effort by Google hasn't produced anything market-ready, yet.
But even if, it was always counterintuitive for autonomous driving to be a driver (heh!) for Uber's future: if there is one hard problem that Uber has arguably solved, it's managing a fleet of drivers and their cars, with all the trouble such a huge workforce of mostly low-income people with low job attachment brings. It is this core competency of Uber that would be rendered worthless by autonomous vehicles.
Instead, car rides in such an environment will be bare-knuckle fight, requiring efficiency in the face of barely-existing margins, and deep pockets to survive phases of intense competition. There is nothing that could stop any business with money to burn to put a few dozen cars into whatever happens to be Uber's most profitable market at any given time and reduce their margins to 0 or less overnight. After all, it doesn't take that many cars to provide service within the not-that-large area people spend their daily lives in, and any number of competitors can create beachheads by convincing a local community to give their app a try. The competitors start $1200 ahead, which translates to two years of free service at the $50/month the typical Uber customer spends.
As long as personal car ownership is still a thing, autonomous driving will actually turn every owner into a competitor. Your car sitting around uselessly during your workday is suddenly just a click away from going to work itself, thanks to the car manufacturer's preinstalled sharing network. They'll take 15% to process payments, but it'll be insured, cleaned after every 5th ride, and back at your office fully charged at the preselected time. Of course the manufacturers will also just let their inventory loose in any area that is potentially profitable. Their capital costs are lower than Uber's, they get the cars at internal prices without dealer markup, they have access first to any new car that people may want to try–oh, and they'll even accept running at a slight loss when they can convince themselves that taking a ride in their new car may convince people to buy it.
Meanwhile, what is Uber but a bro-themed map app with a balance sheet in Ferrari red?
This kind of mindless vitriol and happiness to throw out the baby with the bath water helps nobody, least of which the actual victims of harassment. Then again, I doubt the OP really cared anyways.
And although I do deserve all your scorn, because I am fundamentally certain that current Uber employees are by now co-conspiritors in Kalanick's little frathouse, I am somewhat surprised. Because the comment you're replying to was actually pretty light on accusations of that nature, and was making what I consider a pretty solid case that Uber is simply fucked financially (kinda like the cheap hookers their CXOs frequent when on business in Asia).
Uber's long history of never finding any human too weak to casually stomp into the curb is mostly just contributing to the enjoyment of watching their slow demise.
The first highlight was certainly how the Uber-Randians got the disruption kicked out of them by governments around the world showing them what "monopoly on violence" actually means. Now, we can watch as the paralysed corpse of Uber's upper management watches, helplessly, as Google slowly devours the entrails.
And yes: considering that most any technical employee at this bro-themed sweatshop is (well: "was") employable almost anywhere makes them complicit in everything that went on. It's certainly possible that some Uberdonks didn't notice what was happening around them. But staying with the company is obviously tacit endorsement of everything that has come to light.
I hope they have another two or three years of runway to suffer, at which point it'll be something like a Game of Thrones finale to see all those precious stock options go up in smoke when Uber goes Under. It used to be that people selling their soul to the devil at least were competent enough to make sure their partner was solvent.
Are you having a bit of fun, or are you genuinely passionately against Uber? If you are in earnest, what is your personal connection to the Uber scandals?
If in earnest, it seems a bit out of character from your previously dispassionate posts here on HN. Since you are a frequent poster here, this isn't the first time I have noticed your writing.
Edit: after a second read, I'm finding it pretty unlikely you were being serious. I am not as good at discerning sarcasm on the internet as I should be.
I think I'm actually quite compassionate, and borderline aggressive in many of my comments. Uber always gets that treatment because they somehow managed to position themselves at exactly the confluence of number of seemingly disparate trends I find worrisome. Among them the loss of trust in institutions that are for the most part functioning better than they ever did in the past (media/government/science/etc.), the Ayn Rand-style glorification of individualism to the point of denying any legitimacy to the role of society ("I never voted for the constitution"), and this rather strange backlash against the empowerment of previously disadvantaged groups.
I'm trying to argue these points with the lightest touch possible, especially that last one. But I'm actually quite scared by the changes I've seen in this community, and I'm both at a loss for how it happened and what to do about it, as well as fundamentally convinced that HN is among the places where the next generation finds their ideological bearings, and that it is among the few remaining rays of light that the current leadership in technology has a humanist bend. I'm pretty sure, for example, that Bill Gates' humanitarianism is the result of the culture he experienced in the tech community (because there's really no sign of it in his earlier writings etc).
To think that the next generation could follow not in his footsteps, but in Kalanick's and Thiel's is scary, especially considering the growing power of technology...
Anyway... On HN you'll never find anyone saying black skin makes you immoral, or that jews are secretly plotting...something. It just so happens that this community always finds extremely good reasons to defend this or that fundamental right (free speech, due process, investor protection, whatever) when these powers are used for the betterment of white male cis-STEM-dom. It's definitely not obvious, and I may have a somewhat myopic view. But I'd love to find some statistical method to surface these trends.
According to Crunchbase, Uber has raised $11.56 billion thus far and has $6.6 billion has on hand. That means their total invested over 7 years is $4.96 billion. If their acquisition costs per user to be $1200 as you stated, that means that they only have 4.13 million users.
Given that the company did around 15 million trips on New Years Eve, I would venture that they have way more users than 4.13 million. What are you basing your $1200 per user acquired on?
https://www.recode.net/2016/12/29/14116244/uber-new-years-ev...
At $1200 sunk per user, you believe other companies have an advantage starting later. At what value per user do you think Uber will have an advantage against any late upstarts? $300? $200? $500?
There's likely a reason why the only companies in this space that have survived and remained competitive are all at least 4-5 years old. Are there any TNC upstarts that have done well founded in the past 4 years?
Go-Jek 2011
Lyft 2012 (2007 if you count Zimride)
Ola 2010
Didi Chuxing 2012
Grab 2012
Yandex Taxi 2011
Careem 2012
Uber 2009
99Taxis 2013
EasyTaxi 2011
Technology companies must be willing to disrupt themselves before the competition does. Your argument here is like saying Netflix should not have pursued streaming because it rendered their core competency of DVD-by-mail logistics worthless.
Facebook also disrupted itself by investing heavily in mobile. It's desktop experience is still useful, but it's mobile experience is what really matters today.
The Innovator's Dilemma is very well understood now, which is why you see many of the big tech companies dominating now. Disrupting your own business is part of the DNA of the biggest tech companies today like Google, Facebook, Amazon, Netflix.
Kalanick described autonomous driving as existential to the company. He didn't see autonomous vehicles as a disruption, but as the natural progression. Human drivers are a stopgap measure until the technology is ready. The question is when that will be, and whether Uber can last until then.
Why doesn't anyone ever ask if the competition can last until then as well?
Remember, you don't need to be faster than the tiger, you just need to be faster than whomever you're with when a tiger shows up.
I use Lyft in the bay area a couple of times a week and haven't noticed any gaps from when I used to use Uber.
It may take a few years but there is just too much incentive to move away from the technopoly companies like Uber, Amazon, and Google etc. to decentralized open platforms.
Furthermore, the anonymity from such technologies would greatly undermine trust between riders and drivers. Neither would know if their counterparty is a theif, rapist or simply doesn't have a valid license, safe vehicle or sufficient insurance in case of an accident.
Really, if you stop and think about it, the decentralized model has so many obvious flaws that make it unworkable except maybe in a country with a despotic totalitarian government that uses TNCs for spying on people, such that they are willing to suffer the shortcomings for the benefits of anonymity.
Beyond a handwavy "[there are] many viable paths for solutions worked out", how exactly would you architect such systems so that you can build a solution that produces a customer experience that is competitive against a mostly centralized system now that your solution needs to contend with inter-server latencies that are about two orders of magnitude higher due to the speed of light and the network hops involved?
The essential parts are: finding nearby drivers, maps on two screens, communicating updated position between the two phones, and payment. Finding nearby drivers means that a group of devices in the same geographic area need to be able to talk to each other. I feel like this is right up the alley of distributed systems. The GPS updates really only require communication between the two devices at ordinary internet speeds and updates every few seconds. For payment I do not need to invent anything, we can use Bitcoin, Ethereum, or any of the many high-transaction solutions that are coming out built on those or other cryptocurrencies. So if you want to suppose that it is impossible to have a high transaction rate with cryptocurrency, I don't have time to get into it. But I will say that the main thing that makes people think that is that they assume zero trust and that every single computer must validate every single transaction. Those just aren't _actually_ requirements.
But give it 2 years, maybe 5 years, this stuff will get sorted out. Until then there will always be naysayers saying it is impossible.
Huh, I don't know. There are about three competitors in my area. They all have an app. I enter my destination, someone picks me up, drops me off. Simple stuff. What's so great about Uber exactly?
Uber also have better minor features that you're less likely to appreciate until you need them like lost item recovery and spoitify integration.
17.5% stake in Didi Chuxing is nothing to scoff at, especially considering that Uber has done better than both Facebook and Google as far as China is concerned. Apple is the only US tech company that is really succeeding in China.
33.6% stake in YandexTaxi and international Uber users can continue to use the app to hail Yandex drivers when visiting Russia. That's also far far better than Facebook and Google have fared in Russia.
In India, Uber is beating Ola. In Latam, Uber is easily beating both 99Taxis and EasyTaxi. In the Middle East, Uber is beating Careem. Uber is absolutely dominating in Mexico and other Latam countries.
Besides Didi Chuxing and YandexTaxi, the only international competitor doing well relatively against Uber is Grab, and AFAICT, Grab's success relative to Uber is largely isolated to Grab's home country Malaysia and Indonesia where they have invested most heavily. If you take all of Southeast Asia, it appears to be neck and neck.
Uber is doing well in relatively well in Europe all things considered, and winning in Europe is just a matter of time. I highly doubt the local taxi offerings will survive autonomous vehicles.
What is the competition in Europe besides existing taxi services?
NuBank is the only tech company in Brazil I know of with solid engineers. Globo isn't that bad for frontend. UOL also has one single solid team backend with about four solid engineers.
Brazil's biggest problem is that it doesn't invest enough in education, especially English language skills. Every country competitive in software engineering has relatively strong English as a Second Language skills (at least among their brightest): China, India, South Korea, Japan, Russia, Europe.
If you don't have solid English comprehension among your country's most talented, you'll never produce world class engineering talent in sufficient quantities since the overwhelming majority of technical content is in English.
> Employees who, by now, are self-selected to consider any moves to heal the toxic culture as assaults on themselves.
What do you base this off of?
> What's still ominous is the complete lack of a moat around their business. Every driver and every customer is a free app install away from giving any competitor a shot.
Is this actually true? Uber is the only option in many countries. In the United States, it competes with Lyft... which claims it is on the verge of profitability at the same price points. It appears that the large networks these companies have built do provide a moat. Both Lyft and Uber now employ similar regulatory capture tactics that their taxicab competitors used to prevent competition: from lobbying state governments to override local legislation to conducting public activism campaigns.
> Their autonomous car project is in deep trouble
True, Cruise and Waymo will almost certainly beat them to market. In this regard, they're in the same position as Lyft.
> ... managing a fleet of drivers and their cars... It is this core competency of Uber that would be rendered worthless by autonomous vehicles.
Is that actually how it will happen? Assuming Waymo is first to market: they've teamed up with Avis to perform their fleet management. If any other company releases even within 2 years, is it clear that they will prevent other players from taking market share? How many autonomous vehicles have to be on the road to actually take business from Lyft/Uber? Is it a foregone conclusion that Lyft or Uber won't partner with a rival autonomous company? Do you expect that any manufacturer has the ability or the capital to completely flood any market with autonomous vehicles?
> As long as personal car ownership is still a thing, autonomous driving will actually turn every owner into a competitor.
Is this actually true? Do you expect most people to take the car seats and sports equipment out of their daily driver after they commute to work, and before they come home? I don't think that this is a credible option. Since it is unknowable what owners' behavior would be like, instead look at Getaround's inventory. You will likely see 100 cars for the whole of San Francisco.
Furthermore, you disregard the essential cleaning/inspection that must occur at the end of every shift of driving. That would be extremely unlikely to be provided at 15%.
I get your sentiment, that it just feels like it has these headwinds. But the only tangible point you raised was the fact that Uber still loses money quarter after quarter. The rest of your points were an exercise in making up numbers and speculation without justification.
It's easy to see that Waymo has a big lead, but on what basis are you making the claim that Cruise is ahead. Disengagements? That's still far from being reliable measurement to compare how companies are doing against one another if they aren't all testing under the same conditions. You can easily encounter more disengagement scenarios by testing in more challenging environments.
Uber and Waymo have been testing in challenging environments for a while now. Cruise only recently started testing under more challenging conditions. Their published disengagement figures are all based on 2016 miles, which were "planned tests to avoid unexpected behavior".
Pretty much all the other car companies are non-existent in terms of testing in challenging environments. BMW for example had one disengagement for 638 miles driven in California in 2016, but they only drove 638 miles, so the question is where they tested. The one disengagement they had was on 101 northbound.
You can look at the reports yourself to see how incomparable they are:
https://www.dmv.ca.gov/portal/dmv/detail/vr/autonomous/disen...
https://www.dmv.ca.gov/portal/wcm/connect/6829bbb3-1c28-401e...
In your opinion, who will be the second company to market with an autonomous car, since it looks like Waymo will be first?
Is that L5 in suburbia during good weather?
How much better than humans do self driving cars need to be before they are legally allowed to operate without a human behind the wheel to handle disengagements? Does it need to be better than 50% of humans? 95%? 99%?
How much will legalization be slowed down once we have the first death from a self-driving car without an operator? What about the first time a death happens without an operator and sensor data demonstrates that the company that developed the car is at fault?
If several companies achieve L5 but those L5 vehicles aren't street legal until legislation is passed allowing them to operate without a backup driver, does that count as all those companies taking their L5 to market at the same time?
What happens if L5 cars in good weather are possible, but they aren't good enough in inclement weather? Few people are going default to a service that is 100% autonomous when they aren't reliable the moment the weather changes. What happens when it snows? What happens when you have flooding like in Texas this week?
Autonomous cars might be profitable in developed markets with strong currencies, but how much longer after will they take to become profitable in markets where labor is far cheaper?
Will the deflationary effects of automation in other sectors create enough unemployment and underemployment to drive the cost of labor down enough that it delays profitability with SDCs?
What infrastructure do you need in place to manage self-driving car fleets, keeping them clean and well maintained?
Without the actual presence of another person, the driver, how will we create implied presence that respects privacy and prevents vandalism and damage? What happens when some drunk vomits in an autonomous vehicle?
Google is ahead in SDC technology, but they don't manufacture vehicles. Uber is doing quite well as well, but also doesn't manufacture vehicles? How will manufacturing partnerships evolve to build out a fleet? How quickly will these manufacturers be able to retire a human fleet today with an autonomous fleet tomorrow?
Once there is no longer a driver, how will SDCs cope with other human drivers that are more likely to be aggressive, knowing that the SDCs will err on the side of caution?
Will market demand for private ownership be sufficiently valuable in terms of new car sales revenue, that those with SDC technology find it more valuable to just sell their SDCs on the open market instead of fleets manufactured and operated by the same TNC entity? I suspect Waymo will be first to market, but their requirement to partner with an auto manufacturer will lead them into a negotiation where the auto manufacturer will get to sell their vehicles to whomever they want on the open market. No car manufacturer in their right mind will allow Waymo to retain full control, since they'd be signing their own death sentence if they did so. If autonomous is existential for Uber, it's also existential for the auto manufacturers as well. The only way Waymo retains control is a buyout of one of the major car manufacturers and all the liabilities that entails such as dealerships and maintenance of cars still under warranty for a decade or longer.
How will two SDCs from different manufacturers deal with one another in ambiguous situations?
What will the role of patents play in this market? An L5 car is likely going to require so many patents that each of the top major players in the SDC space will own essential patents, making cross licensing unavoidable. Uber will end up licensing from Waymo and Waymo will end up licensing from Uber.
From an L5 safety perspective, every car on the road, human-operated or autonomous is a liability. This means that every SDC manufacturer has a vested interested in supporting their competition in being safer. If Uber helps make Waymo vehicles safer, they also increase the safety for their own customers. The same applies to Waymo helping Uber vehicles become safer.
Basically, there are so many things that can and will go wrong to slow down the adoption of SDCs and make adoption geographically uneven, that I suspect that it's going to be a sufficiently gradual change that being first to market is going to be relatively unimportant. SDCs are evolutionary, not revolutionary, thus the incumbent will greatly be favored. The cost savings of gradual SDC adoption will be spread across all rides as quickly as it makes economic sense, but no quicker. Whoever has the strongest relationship with the most riders will win out in the end.
For people to give up car ownership, transportation needs to function with the reliability and confidence of a utility like gas/water/electricity. We're going to have a hybrid models for years before fleets become entirely autonomous. Four nines of reliability is non-trivial and that's what you need to be a utility. L5 only satisfies some of the requirements necessary for a four nines transportation service.
SDCs will be on our roads far sooner than the naysayers predict, but they will take much longer to replace humans than those bullish on SDCs predict.
That all said, if I were a betting man, I'd say Uber will almost certainly be "second to market", whatever your definition of "to market" is. Why? It's ultimately going to come down to a race for data and Waymo and Uber are going to do much better than the rest of the industry in terms of data collection and best in class infrastructure to crunch and manage all the data collected and to run simulations. Like Apple, Google and Amazon, software companies will have an easier time becoming hardware companies, than the hardware companies will become software companies.
Lastly, I may be totally wrong about everything above. If someone has contrasting viewpoints about how the general transportation market could switch to SDCs suddenly instead of gradually, I'm all ears.
Probably the only scenario that leads to a sudden influx of autonomous vehicles would be intense competition that pushes all competitors' break-even point to many years. Another necessary condition would be extremely low-cost retrofitting for existing vehicles. Still, I can't imagine a single scenario that leads to developing countries adopting autonomous vehicles within several years of them being available in industrialized countries.