Thank you Brian.

This is just like Enron. People are talking about and valuing things on technology that doesn't even exist (super duper amazing routing algorithms that will do jack **** for efficiency over what we have already, autonomous cars that don't exist yet and that Uber will be the last player to crack). All topped with a crown of abstract buzzwords like "logistics" that don't actually mean anything.

It's Enron all over again. They weren't just a holder of fairly boring utility companies. Oh no, they were a dynamic asset pricer, newfangled derivatives market maker (creating markets no one had every seen before!), inventor of amazing mysterious software to price assets differently and skim the new efficiency. They were doing incredible, mind blowing things with their 20,000 employees (what a moat!) that the "old money" couldn't even comprehend, and setting the world on fire with their brilliance; tens of trillions was the top.

Sane people bought into this bull**** by the millions. It's ridiculous. Uber are a global hands-off taxi app company who undercut the current legally protected cartel by flouting the laws - both labor laws and taxi laws, and have established a nice little niche (as yet unprofitable) doing it. That's it.

They'll capture a fraction of the $100 global billion taxi market with increasing difficulty (the low hanging fruit already gone), maybe make the market a bit larger, and establish a reliable presence in some areas.

In five years or ten, autonomous cars will come along (NOT developed by Uber), and both big and small money will be behind this revolution, in fierce competition. The driver payment differential will disappear, destroying Uber's advantage; the legality differential will disappear, destroying Uber's advantage; large fleets will appear from people like GM, and Avis, and taxi companies, with existing local infrastructure for storage and compliance, as well as the big tech companies like Apple and Google. Everyone will fight for survival in this new lucrative world. It will be a cut-throat business with the tiniest of margins; the differential will be in fleet choice and buying power, effective warehousing and fleet management such as charging infrastructure, not routing. Uber is very likely dead in the water here; they won't be first to autonomous cars, and will never have the revenues or infrastructure or available profit to burn to compete with the big players on a global roll out of vehicles and support infrastructure.

BrianTheMick -

So...are you willing to give us an opinion on how many companies have the IT assets to replicate the Uber app at its target scale?

Is n = 50? 500? The # of taxi companies that could hire a comp sci undergrad for a weekend? What do you think?

Mihkel05 -

Yeah, a lot of interesting challenges, definitely hoping to transition into a DevOps role one way or another.

What do you think about this...

Sub,

Yeah, I mean we're talking about real time updating of a ranking system instead of running batch scripts. They spent like 15 years running ****ing batch scripts to update their search engine. Letting that sink in for a while to understand the complexity of just getting something to update smoothly at that scale is insane.

I think my friend is actually actively being recruited for a director spot in exactly what you wanna do. Lucky for you he is too busy making the world a better place via some other piece of software.

Brian,

You have posted several outright stupid things. That was the dumbest yet. This is like claiming Amazon has a simple ecommerce site. "You just click a button and buy stuff" without the amazing logistic work, company culture, whiz bang recommender systems, etc

These things you think are normal or trivial, are just because you don't understand the subject well enough to appreciate how difficult the challenges are and how novel the innovation is.

Weirdly, all the tech people in this thread have the exact opposite view of the people who lack the technical skills to comment on a tech company.

Mikhel's link above claims that Uber just put their first test car on the road. How far behind are they? Google's first public test car came out four years ago. Uber are so far behind it's ridiculous.
http://www.macworld.com/article/2925...est-drive.html

Bill Gates seems to say a lot of stupid stuff. Uber is way, way behind. They wanted to buy Nokia's HERE mapping (the best in the world) but lacked the money; instead it was bought recently by a conglomerate of car makers with hundreds of billions in assets.

Meanwhile, Apple (with $200 billion cash on the books), has been going on a hiring spree hiring the best in autonomous vehicle research.

You think Uber with its 2 billion in startup capital - most of which it needs for operating expenses - and starting many years behind, is going to get there first? That's just a ridiculous notion, frankly.

Even if they solve it first - at best they would beat others by six months - they still have to make the ****ing car with sufficient volume and capital outlay to start making inroads on human transport. That's $100 billion in investment right there (look at the books of car makers, for example). Then they have to set up the infrastructure in every city to support it (charging, parking, maintenance and cleaning) vs someone like Avis with infrastructure already, all over the world, who simply has to buy the cars. You think they are going to do it, rather than car makers? Or Apple? Or large fleets?

The whole thing is so ****ing absurd. Starting a tiny autonomous vehicle research unit is probably worth it just for the extra money they will get from the buzz that idiots and investors will buy into. I wouldn't be surprised if that's why they're doing it.

It started in early February partnering with Carnegie Mellon.

You are so stupid and dishonest it is impossible to engage in discourse. You spout lies and nonsense continually.

Haha, awesome. But actually I think it's pretty clear why Gates has this opinion---and I'll be surprised if Uber isn't the first to self-driving cars:

1. Unlike almost everyone else, Uber's incentives are completely aligned with the goal of autonomy.
2. They are currently and will outspend everyone on R&D. (As Gates notes in the Financial Times interview.)
3. They have the potential to mine data assets on a completely different scale than their competitors.

[1] Analogy to Tesla, with its rounding-error assets and R&D budget. Didn't matter, obviously Tesla was going to be the first to a compelling BEV, since that's its mission in life and the incumbents have (at least moderately) strong disincentives to making BEVs.

If anything, the ICE incumbents have even stronger disincentives to autonomy, so seems pretty likely the same pattern will play out here. Google's moonshot and Apple's program do have aligned incentives, but they're hardly facing the life-or-death stakes that Uber's playing for.

[2] Everyone knows what's at stake here, and Uber can and will raise any amount of money it can find a use for. Representative example:

[3] Today when discussing Tesla's first baby step toward autonomy, Elon said:

Suppose Uber develops a cheap external hardware suite for autonomy-learning and pays x% of their drivers to use it. They could make Tesla's fleet (much less Google's) look data-starved by comparison.

Is there a video of what Gates is supposed to have said? Is seems a bizarre comment to make given the tone of the interview. The Financial Times transcript of the interview has no record of him saying that. Have a read of the official transcript:

http://www.ft.com/cms/s/0/ed5ec9c4-1...b03d71480.html

Here is what is quoted:
He says the "real rubicon" in the interview, which is in the transcript, but there is nothing of the latter in the official transcript.

The claim that he said this about Uber - and their research budget - seems to be based on a tweet by a reporter at the event. It seems a really weird thing to say; Gates is a careful man and doesn't pick winners. Nor would he know how the multi-billion research budget (which include a large autonomous car budget) is split up. Is it possible she simply misheard? That she lied?

Do you have any any evidence to support your assertion that Uber is currently spending more than anyone on autonomous car research? It seems quite unlikely given that they just put a single car on the road a couple of months ago.

I'll get to the rest of your post if we are actually talking about facts.

And yes, what you've said is completely absurd. I'll explain why in a little more detail if my last post didn't lay the bread crumbs thickly enough.

I don't have access to the transcript, but I would think Fortune would have issued a retractment by now if someone randomly invented that article out of whole cloth...

And yes, additional evidence such as buying 40-odd top Carnegie Mellon scientists (including a team of 20 who had already lined up tens of millions in VC capital for their own startup), and their overall hiring rates, I think it's extremely likely Uber is spending more than anyone on autonomous cars.

This is closely related to the incentives problem; if Mercedes is truly serious about autonomy, why is Tesla again two steps ahead of them with its teeny weeny rounding-error resources and thoroughly poached staff?

P.S. Is Gates a "careful man" or someone who "seems to say a lot of stupid stuff"?

When you're quoting him, he seems to say stupid stuff.

Here is the part of the actual transcript since it's behind a paywall. The tone of the interview is about technology changing labor laws:



Does it sounds like he says Uber has larger R&D? That it's "primed to take the lead"? I suspect the quote is simply misheard or false.

By the way, he's one of the people who's spot on about solar, which is further proof that Musk is an airhead on the subject (or lying, given that he owns a solar company).

Anyway, do you have any evidence besides a tweet of a quote by Gates that Uber is spending more than the hundreds of millions to billions that others are? It seems frankly extremely likely to be false. Hiring a few dozens of researchers years behind others, combined with how little money they have, seems to point quite strongly to them being a joke in this space.

Ok, I'm willing to assume Ms. Kaminska had mixed her meds that evening, so we'll dismiss that article.

Haha, staying away from the EM digs after you made me feel like a sad little fanboi. (But srsly, read the bio!)

Well, although Uber hired the Carnegie Mellon scientists recently, that clearly doesn't mean their research is ab initio...again, the raid included people like John Bares who "...led [the National Robotics Engineering Center's] largest DoD unmanned vehicle design programs [from 1997-2010]", as well as another team of 20 who had already lined up tens of millions of VC funding before Bares recruited them to Uber.

Stepping back, the reason I think Uber is probably spending more on autonomous driving is that it's run by people with unlimited cash, who really want to win, and can see the writing on the wall as well as any of us ITT.

I'll agree that they need this. But no more than the car makers with hundreds of billions in collective assets, and R&D budgets in the high single digit bililions (each), who'll be left holding the bag if other car makers get here well ahead of them.

I mean, this is chain of events for Uber to profit from their research on self driving cars:

1. Uber beats all other researchers while starting a lot further behind and with a lot less money. They'll also beat the incrementally improving car makers (peak hour traffic -> highway -> all streets) to get out the first fully self driving car prototype.
2. This is then put into production (by who?) and built in sufficient volume in a short enough time that no one else catches up
3. Rather than sell to the huge market demanding these cars, Uber uses these cars for its own use, to replace its drivers.
4. None of the people spending billions of dollars on this apparently not that difficult problem catches up with autonomous car production in the 2-3 years it takes Uber to ramp up production of these cars, replace its drivers with these, install bases all around the world for storage and charging and maintenance, etc, large enough to capture a significant portion of the transport market (i.e. at least 50 million cars, or 10x the yearly output of Ford).
5. Consequently, Uber captures a signficant percentage of worldwide autonomous taxi traffic (at least 10%), justifying its current valuation.

You're a very smart guy. Surely you realize how absurd this chain of events is? Their research adds zero value and is just a gimmick. Or perhaps they think like you do, boundaryless and trying anything, without actually thinking this through. If someone as smart as you can do it, they can too.

I agree that a) autonomous cars will change the world and b) whoever gets a decent share of autonomous worldwide taxi revenue will do very well with enormous revenues.

But Uber's research in this space isn't worth a single cent. It's worthless.

What's really going to happen - and where Uber has a shot - is this:

- Car makers eventually come up with fully autonomous cars. Most will get there at the same time, due to increasing research dollars, the hardware finally being ready, and corporate espionage.

- They start rolling off the lot, and Uber, which has managed to maintain a decent market share in the human-driven or carpooling market, either buys these up into fleets, or starts using the autonomous capabilities of third party owners who it pays a cut.

- Intense competition enters the space; everyone including very big money tries to get in on this space; the first cars sell for a premium.

- Uber's app installed base and brand awareness, with a large stable of habitual consumers leads to a seamless transition to an increasing number of autonomous cars.

- As a consequence, Uber captures reasonable market share.

That's how it's going to go down if they're going to make it. Their routing algorithms, their autonomous car research, it's all bull****. They're an Internet app maker and the market share they can maintain among their app users is their only competitive advantage. Their future prospects depend entirely on how well they run their taxi service between now and then (retaining driver networks, having good customer service) and how aggressively they take over local taxi markets and eat into commuting with things like ride sharing.

Which is why it's nonsensical to talk about their autonomous car research; it's utterly irrelevant. The only thing that matters is how much they can disrupt and grab market share before autonomous cars come online.

TS -

Those two posts teased out some hidden assumptions I'm operating under, but I'm still not sure they're absurd! Will bb tomorrow.

You make some great points but still underestimate their software (and I am not some blind Uber supporter -- I have already stated ITT they don't pay their drivers enough to make it worth their while). Also, in cities where you say you can get a cab in less than 5 minutes, I agree as they already have dedicated cab lines or enough cabs going by you don't need to think about it. However, Uber provides a superior service in the majority of cities as most cities (even most major cities) have a much longer wait for cabs. It is also my argument that the cab companies will invest to close the service gap in these cities and are probably already doing so. We will see who wins out.

Don't know the cost, but taxi companies in Las Vegas (and I assume elsewhere) already have internal computer dispatch systems that give rides to drivers based on distance from the customer using gps.

They also have an app called Curb that is across taxi companies. I think upwards of 5 taxi drivers use it.

The nice thing is that google maps and/or waze (and/or ANY existing mapping/routing online app) can be used by developers to do most of the hard work. The customer database and credit card processing parts can be purchased off the shelf.

FWIW, this is decidedly untrue.

Not to derail this thread. But how soon do the experts ITT think self driving cars will be common? Inside of 20 years?

It ranges from 3 to 20 years, with many saying 5. They'll obviously happen first in places like San Franscisco where there's ultra high resolution mapping and tech and the city/roads department is forward thinking enough to make everything autonomous friendly. The first cars will be hacky (using high resolution maps + GPS for positioning, probably some kind of signalling installed/carried by workers and police when there are road works or changes, possibly car-to-car signaling) rather than truly autonomous and navigating without aids. You can do hacky autonomy 3-15 years before true autonomy is possible.

Atleast you've stopped spouting off that idiocy about higher resolution cameras. Maybe you can learn some basics about a subject before commenting!

Not really. There are a few people who know nothing about the subject but have commented who may serve in some PM role or junior level interface role, but I have not seen anything compelling to by anyone who has any experience and any familiarity with Uber to have a different opinion.

Or perhaps you think their software is literally work 50-500k and just a glue layer for several different APIs.

That is literally the view that is being put forth by Brian/TS and others. Where are you confused?

Cost != valuation.

You continue to ignore what is arguably the most important factor, which is route density.

TS -

You jarred so many thoughts loose, I need one more day to fit them all in a narrative...

BrianTheMick2 -

Doubtless there are some very nice Platform-as-a-Service offerings that hugely simplify building an Uber-like app. I'm thinking about scale here.

Anyways, to seem less coy, I'll estimate:

• There are ~300 companies with the IT assets (infrastructure+operations+appdev) to deploy a car-for-hire app at Uber's current scale.
• There are ~25 companies that could handle Uber's target scale. (The blue chip Internet-scale businesses, a few multinational banks, maybe the biggest airlines, perhaps financial backbone processors like DTCC.)

You are making a small mistake here. You don't look at current capacity, else you'd find that Uber doesn't have the capacity to reach its target scale.

Uber has the problem of being the first mover here. The amounts it is spending in every jurisdiction it wants to operate in to get regulations changed to allow it to operate (and on lawsuits where it hasn't) are non-recoverable losses that only it is bearing. The amounts it is/will be spending on fending off employee/contractor status rulings/lawsuits... The amount it is spending on obtaining customers will be higher for them than a second/third/tenth competitor.

They really don't have to worry about the big boys until all of the regulatory issues are settled. Publicly traded companies can't take the risk of executives being jailed, etc.