Although Brian I'd argue that their "value proposition" is that they get ahead of others in being the first to have large networks of drivers at all hours, who are reliable and cheap and suitable for nearly any use.

San Francisco is instructive I think - huge revenues, ownership of the space, a young tech savvy populace some of whom are even giving up cars and muni (thus creating their own market).

Still, car costs are mostly miles + driver and given the driver problem I'm not sure they're going to beat the costs of personal transport sufficiently well to be truly disruptive until autonomous cars come along (at which point they're probably screwed).

I dunno, I don't have any opinion about their business model, but am pretty curious about the sophistication of the backend. (I've thought about applying to work there in a 8-12 months after I fill in some missing skills.)

I just think it's pretty hard to look in from the outside and know how much scale is complicating the problem. (Unless you have a lot of experience building distributed applications, perhaps.)

I mean, a smart person can understand PageRank in an hour, but building a Internet-scale implementation looks something like building a 20,000-node Hadoop cluster...

Subfallen,
I have a taxi company. Let's call it Duber.
Let's say you've got 100 Duber drivers circling San Francisco and the built up bay area. You track them via GPS from a central computer, which is trivial, and users request trips via an app you own, also trivial.

So what is the difference in efficiency and price and user experience between:

1) A simple "pick up the closest passenger as they request" algorithm with a tiny bit of add on code to prioritize those that have waited too long
2) Existing taxi routing software of various flavors
3) Perfect software that fully solves the traveling salesman in three dimensions in nanoseconds

Now make that 500 drivers, or 2000.

I don't know anything at all about routing, but I imagine the first 5% of the work is 80-90% of the efficiency/problem solved. What do you think? What's the difference in occupancy rates/mile and wait times between 1, 2 and 3?

No, their "app" as you call it has revolutionized the industry and is much more complex than a simple GPS app. The main reason it did not exist before is because taxi monopolies did not have enough reason to invest in making one, as just putting people on hold and making them wait up to 30 minutes for a taxi to arrive was considered standard practice and consumers had no other choice. The app is not nearly as easy to write as you make it seem, and there really was no excuse for taxi companies not to develop their own, other than they had a monopoly so didn't have to.

Now that Uber is a threat to taxi companies I expect them to come up with similar apps/technology and start providing much better service, which will be a blow to Uber but a benefit to the consumer (probably still several months to a year or more away -- but if you really think a college student can write this in a few days they should do so and sell their version to every taxi company, you could easily sell this as a skin to every cab company across the country/world who don't have enough resources to develop their own). Whoever does a better job going forward will win out, right now that is Uber.

This is just wrong. I've traveled widely and there are many cities where you can order a taxi and have it at your door in under 5 minutes at any time of the day or night. Where that's not the case, it's not because of lack of an app or because of routing problems that Uber has solved that others haven't.

You're acting like advanced vehicle routing software doesn't exist (hint: it does)

Uber's advantage is not what you think. It's three things:

- Their app uses an existing account and your phone's GPS, such that you get pickup with a few button presses

- They ignore regulations while their competitors are bound by them; where others pay up to a million dollars for a taxi medallion, Uber drivers pay nothing; where taxi companies have payroll costs and major compliance costs, Uber drivers are merely background checked, then work for a pittance using their own vehicles in an unsustainable way. This means far lower prices. Nearly all of their price advantage comes from this fact, not routing.

- Scarcity of taxi medallions in many areas means that there are only a certain number of taxis allowed to be on the road, which is often insufficient. Drivers who flout these laws can work busy times and fill the gap, or oversupply an area for fast pickup (given that drivers often want money now rather than money later, even if they're working for nothing in the longer run).

That's it. There isn't some super duper magical VRP solving algorithm that means you get pickup way faster and far cheaper. I can get a taxi to my door in <5 minutes, reliably, every time, where I live. One of the reasons Uber and Lyft did so well in San Francisco is because there was an enormous taxi shortage caused by artificially and perversely limiting the supply to far too few medallions. That's just not true in many parts of the world.


You're completely misattributing the cause of their advantages. It's not their routing.

And somehow people think my characterization of the situation is insane?

He really is the same as steelhouse.

With little/no cost of switching for the drivers and customers, this is the same value proposition one would use to evaluate MySpace.

That is exactly how they can win.

Uber isn't bearing any of the miles/driver costs. So long as there exist financially illiterate people willing to drive for them, they are viable. The supply of such people is more than sufficient and will remain so. The only risk to that part of the model is the DOL.

Why MySpace and not SnapChat? Or Tinder? I think most of their value proposition is that apps for performing everyday tasks on your phone are sticky, and early mindshare and phoneshare and network size tends to determine the winner. MySpace was around in the age of desktop computers, not phones, and was not required for daily tasks.

What I mean is that they need to be lower cost than individual car ownership if they want to majorly disrupt personal transportation worldwide, as is the goal to get above their already absurd valuation. The problem of needing a driver makes that near impossible I would say. At present they're picking the low hanging fruit of a pool of idiots willing to drive for a pittance and bad conditions, willing to devalue their cars in return for immediate cash (essentially a car equity loan), but I don't think this pool is large enough or enduring enough to sustain major growth.

There isn't any evidence of stickiness.

In San Fran, they already are pricing rides lower than the total cost of ownership.

I disagree with the depth of the pool.

Full time Uber drivers in Manhattan make 30-40 per hour on average from what they've told me.

So it takes a full time driver a week to generate enough gross cover the fixed costs of depreciation. That's honestly pretty darn good compared to most businesses.

I think it's a matter of time before drivers get together to share cars and negotiate better maintenance contracts with local shops to drive the costs of the cars down even further.

Something not publicized very much is Uber drivers get to rate their customers. That particular piece of information has been spreading in Manhattan and more and more people are tipping contrary to Uber's recommendations.

http://www.msn.com/en-us/news/us/a-p...ade/ar-BBjuRkA

Yeap, sounds about right. I heard similar stories in Taipei, Phoenix, and elsewhere. That uber driver income correlates with local income and population density is not an accident.

PS: the 90k claim specifically refers to NYC. Uber declined to provide data on the other 60 cities, according to the article your article links to.

Internal letter from Kalanick reporting close to 1 million rides per day in China with plans to launch in 50 more cities in the coming year:

http://techcrunch.com/2015/06/11/ube.../#.8x4hv1:uv12

Many service industries have sustained themselves despite paying lower wages than Uber. Compare driving to say being a cashier; there's a benefit to setting your own hours at a relatively easy job.

25% of drivers are reportedly over the age of 50. I know older folks who drive on occasion and like it. They're not overly concerned with the wear and tear on their cars because they're not putting a lot of miles on them in their daily lives, and a comment I often hear is "Well I wasn't doing anything any way so why not make a few bucks?" They also enjoy interacting with passengers.

Uber is of course monitoring driver signups, retention and feedback. I think it's likely that they'll figure out ways to incentivize drivers should the numbers start falling, and unlikely that they'll fail because they can't get enough drivers (though they may hit a ceiling which may limit growth).

http://www.businessinsider.com/uber-...-report-2015-1

Well i don't know much about the other concepts of uber then uber-pop.

But i do worry about the safety of their drivers, and there where/are many incidents in belgium, france, and the netherlands and if you take a look at all the articles one thing stands out all the protection uber gives is they would pay the fine.

But nothing about protecting their drivers against the other legal taxi drivers, and i can tell you that they are not known as the most civilized group of people if you are taking bread from their mouths so to say, and that is exactly what uber-pop drivers are doing.

And i know there is not much they can do about it, but they are responsible for enabling those uber drivers to break the law[since in these country's they are forbidden by law], and violent and threatening and intimidating situations already occurred in the last year, with thank god no serious injuries.

But i think uber bring people in danger by letting people use their app and have no way of protecting them, and i guarantee you that here in the netherlands[since i can see and hear it here first hand] there are going to happen bad things, and we had a taxi-war here in the netherlands in the past, and no deaths occurred but there where close calls, and that where taxi-drivers among each-others and both had legal standing, and a uber driver is illegal, not a excuse but taxi-drivers are a protective bunch and as i said not all of the bookkeepers kind and they can feel more "ïn their right" by knowing the uber driver is breaking the law, and then you have mob-mentality that can bring on violent behavior, so i fear for any uber driver that get caught by a bunch of taxi-drivers, because they are looking out for the uber drivers and are all in contact with each-other, and following with 2 or 3 taxis and videoing uber drivers are happening regularly, and these are just the ones that get in the news/newspapers.

And i heard talks already among truck drivers, that uber better not try anything in their sector[illegal like uber-pop then i mean] because they will protect their lively hood, and since i have many contacts in the trucking world because of my 25 years in the import-export business and see many owners of truck-company's still on a regular basis, i can tell you not much truck-drivers will get up against these guys they are outright dangerous if they get stolen from like uber-pop is from taxi-drivers, because most of them[95%] are honest hard working but hard men that have worked 16-18 hours days for many years too build there company's and may god have mercy on the souls of the men or company's that threatens their company's in an illegal way, because i can assure you they won't and their personnel won't either, but this is all not current and might never happen, but still.

No i predict that uber is/going to leave a lot of damaged and/or hurt[financially only i hope] in many country's they are starting businesses with-out being completely or at all legal like they are doing now, and they willingly and knowingly put people in potentially dangerous situations that can financially, mentally and even physically or even worse may god prevent, and i don't care how they call it or justify it, they are doing it for money!!!! that is the bottom-line, and i hope no person gets hurt or worse, but if they do i hope the owners of uber get punished to the full extent of the law and i say also beyond.

Because every business they are threatening/competing with now have started with making sure they have all the licences/permits and so on[we have too assume], and now uber come's along and says we don't need all those thing we just start a app and some paperwork office and take money/business away from company's that [suppose to] did it all right? No that is not right in my eyes it is arrogant, pretentious and i would even call it theft, and if they think they can do it cheaper, better, more effective? ok but pay your dues [if they are fair or not, not the issue] and prove it the right way, because just doing it and see if you can change it that way, and even [easily foreseeable] bring people in danger doing it? then you have sociopathic/psychopathic tendencies or at least totally unscrupulous, and are a other fine example of the today's people leading the global business-world.

A good McDonalds cashier makes well over $100k/year using Uber's methodology.

Is Uber against the law in the Netherlands?

He has no idea what he is talking about. He is rambling about one of their 5 products offered in Amsterdam, and then rambles about how truck drivers are a violent and sordid lot who won't tolerate someone better at their job taking it away.

http://www.macworld.com/article/2925...est-drive.html

Obviously most people in this thread have no understanding of what Uber is doing, and just ramble at length. Or make snarky comments only to reveal an underlying thesis that is ridiculously comical (BoredSocial in particular).

Uber isn't interested in a sustainable driver model and doesn't care at all.

Care to elaborate?

+1 to the trucking industry being a rough place. Thankfully for the 'uber for freight' people (whoever they might be) they won't be hurting trucking companies.

Dear god if someone did to European trucking companies in Chicago what Uber has done to taxi drivers there would be fire bombings.

I think we're a very long way from driverless cars in places like NYC, San Francisco, Shanghai and Mumbai. I can see it happening sooner in places like Phoenix, but that's not where the money is for Uber.

I mean, unless you gave me something like Uber's data assets to run simulations against, I can't really say how 1), 2), and 3) would compare at Uber-scale.

And even then, I would have almost no insight into the operational challenge of doing any of the three at Uber-scale!

There really are serious difficulties with scaling backend services (even "simple" services like file hosting), which I'm hoping to understand better as I give myself a crash course in distributed systems over the 6 months.

But, take this example. I'm currently doing systems programming for a F500 that has one of the 20 largest software revenue streams---so we have plenty of IT assets.

Recently we made a seemingly very spewy ~$500M acquisition of a Company X that is growing but losing money hand over fist. (Company X sells subscriptions to a web app that does project management; there are many close substitutes for Company X's product.)

A big part of the internal justification we were given was that X's platform can scale better than anything we've been able to build.

That's a really good point.

Just trying to imagine driverless cars in India or even Malaysia. They'd be forever stuck at every intersection/broken traffic light/place where you have to break the law and push in just to get by. And where you have to follow far too closely for safety just to stop people taking your spot. Driving in peak hour in many Asian cities requires total abandoning of road rules, and taking whatever illegal opening you can such that others with legal right of way have to yield or smash into you.

Then you have the kidnappings. A driverless car, built for safety, merely needs to have someone stand in front and behind and it's unable to move, and you can help yourself to whatever occupants you want. The only way to stop this is to program it run to over humans, which requires human-like context awareness.

Uber-pop is[hanging appeal that is, but has no chance] and the other methods are not that much cheaper then normal taxis and proven to pay less then normal taxi-drivers get payed, and the uber drivers are wearing their own car down instead of the bosses one.

And i did mention several times i was talking about uber-pop, and since uber-pop i forbidden in the whole!! of the netherlands and the app is! used in more then only amsterdam so that like a cult-like follower/defender of uber mihkel05 is talking from the wrong part of his body again and only wants to hear his own "truth" again!!!, he sure gets old soon doesn't he?

Another insightful comment. We have "I don't think Uber will be disrupting long haul freight" and "Truckers are violent" so far.

Really backing up those comments to show you know anything about this.

That is one of the issues that their company is facing. If SDC take 20y, they are almost certainly ****ed. However, considering they are already testing their own cars, I'd say they have a much better understanding of their own internal finances and projections then us just speculating. Regardless, it is a known risk.

How long ago did Google manage to stop running batch scripts for indexing? Few years?

They developed NewSQL before they stopped running batch scripts? Around the same time?

I dunno. But luckily we have TS here to set us all straight on how these algorithm scale effortlessly and there is no operational advantage to Uber's technology. 50b to flag a taxi. God his thesis is so amazing.

The advanced "algorithm" they are currently using is: drivers drive around willy nilly, rider requests ride and places a pin on Google maps to let the app (and later the driver) know where he is, app pings driver that is closest to rider and he is directed to go where the pin was placed on Google maps, once driver has arrived and the rider is in the car he indicates to the app to start charging the rider and if destination is entered google maps gives turn by turn directions to direct the driver so he doesn't get lost.

They do have a pretty smart surge pricing calculator, but they are relying on the drivers to use their wetware to determine where to seek a ride.

It is important to note that the same technology can be used to direct a driver to go obtain a burrito from the local burrito shop and deliver it to a customer.