Hacker News Reader

> Why on earth would autonomous taxis be a monopoly?

I'll bite. First, there's the entry point. If it's a taxi, you as the rider don't own it, which means you have to summon it. The most ubiquitous company here has a significant advantage, because they own the app or phone number that you have on your phone. You won't have to learn a new trick every time you go to a new city. It's similar to Google's monopoly on search.

Being driven around is also somewhat of a commodity. People don't have strong preferences that can't all be satisfied by a single company. You want a nice car or a cheap one? A solo ride or a group one? Easy enough to do in a single app. Whereas an industry resistant to monopolies — say restaurants or productivity apps — are subject to a variety preferences and whims.

There are also economies of scale. Major players like Uber and Lyft have the marketing budget to drive far more growth than competitors. They have the budget to invest in the latest technologies. They have the budget to outright copy or buy competitors, like you see Facebook doing. They have the budget to form partnerships and integrations with other big and powerful companies.

Network effects can play into this as well. Pretty much everyone has Uber and Lyft, which enables things like ride-sharing (Uber Pool and Lyft Line), or splitting the cost of rides with friends, etc.

This is all just off the top of my head. I'm sure others who are better versed in business than I am can think of more reasons for why autonomous taxis are well-suited to monopoly, and counter-arguments as well. But it's certainly not a ridiculous presumption.

Literally every single one of the dynamics you mention is already true for taxis driven by people.

Autonomous taxis are taxis, which we understand, with computers replacing the people. That's all they are.

Magical thinking abounds.

Not true.

First, the taxi companies never made apps, so the lock-in factor just wasn't there.

Second, the absence of an app makes it hard for any single company to succeed with every option that customers want. Willing to pay more for a nicer car and a professional driver? Or want to save money by sharing a ride with someone else? Sorry, you get whatever cab is closest, unless you're willing to look through the phone book, and the cab companies can't stop you from seeing their competitors in the phone book. With an app, I can get all of my options taken care of without ever seeing another company.

Third, taxis are heavily regulated and even have to buy medallions from the limited share the city makes available. Uber and Lyft completely skirted regulations, removing an artificial barrier to monopolization.

Fourth, all of the above greatly diminish the effects of the economies of scale that Uber and Lyft enjoy. Why spend a bunch of money on, say, advertising, if it isn't going to help because people simply jump in the nearest cab regardless? How can you block the competition by copying their features if nobody has apps with features to copy? How can you buy out the competition if you don't have enough cash because needing to invest in medallions and drivers slows your growth?

It's not an accident that Uber and Lyft have put companies like SF's Yellow Cab out of business.

With self-driving technology, the economies of scale become even more influential. Who's going to get first dibs on exclusive partnerships with car manufacturers and tech companies? Who's going to have the funds and the clout and the experience to negotiate with regulators and governments? Who's going to have the software engineers to develop their own self-driving technology? Not small taxi companies, that's for sure.

> the taxi companies never made apps

? https://www.google.pt/search?q=itunes+taxi+app

I'm talking about before the Uber-Lyft duopoly came into being. It's obviously futile to make an app now. How many taxi companies had mobile apps before 2009?

Not true at all. I can't hail a cab with ease from 90% of the places I might want to call an Uber. They also can't be counted on to deliver a quality experience. Therefore, people are switching to the new technology of rideshare apps. Where the incumbents have a huge advantage because using them if you have already is seamless. Self-driving is just the next step in the chain.

Right. The idea of app based hailing is far more impactful on the monopoly question than the self driving part.

And Uber's delusional investment premise notwithstanding, it's pretty clear that there's not going to be a genuine monopoly in app based taxi services. At minimum there will be a few major players.

But the relevant point is that there's no reason to think autonomous driving will strongly affect the monopoly dynamics of taxi services really at all.

I agree with you to be cautious not to assume either of these with certainty, but just to play devil's advocate:

Assuming that self-driving car tech gets to the point where it's viable as a business, why _wouldn't_ you assume that it'd be huge? Trucking alone is a massive ~$700 billion/year industry in the US (https://www.selectusa.gov/logistics-and-transportation-indus...). Even if you gain only a few % points of that for autonomous vehicles you're sitting on top of billions in revenue. The auto industry adds another ~$600 billion/yr on top of that. Sounds pretty big to me, and there's lots of room to grow from there.

Now, as for monopolies in the market, I don't buy it quite as easily, but I can see it due to the usual "rich get richer" effect. Not to mention training data -- any self-driving company that starts to serve more rides will be accumulating more training data, allowing them to improve their product more and making it that much more difficult for second-in-lines. Look at Wal-Mart, where competition requires big infrastructure, they're bigger in market cap than their closest 5-10 competitors put together.

Switching costs for consumers is low. There is really no difference where I live between Uber and Lyft, and I could see another two or three competitors doing the exact same thing.

Another industry that has low switching costs for transportation are the airlines. They had to implement rewards programs to keep customers loyal, and that still hasn't stopped the race to the bottom in terms of price.

The software and training data set an extremely high barrier to competition.

Once you've solved self-driving taxis, you've also solved a large chunk of self-driving commercial transportation. AI trucks could kill off rail, in addition to manned semis.

From there, it's a small leap to becoming the backbone of an ultra-efficient on-demand national distribution network.

All without the massive operating overhead of human labor... No sleep, no breaks, no weekends, no healthcare, no HR, no wages, just the price of electricity and the low/infrequent maintenance costs of electric vehicles.

Eventually. I think the taxis are 5 years away from 24/7 availability in NYC, maybe 3 for San Francisco, and the grand vision is probably 10 years out.

But... Yeah... This is going to be huge.

> I think the taxis are 5 years away from 24/7 availability in NYC, maybe 3 for San Francisco

lol

That's another amusing point of view I see here a lot. The use case of taxis in a major city is by far the most challenging use case for autonomous driving. Mark your calendars for 5-10 years after you first see an actual autonomous vehicle having any type of real world application at all. Which you haven't.

The whole idea that taxis are the vanguard of autonomous driving is obvious patent nonsense, why it persists remains a mystery to me.

> Mark your calendars for 5-10 years after you first see an actual autonomous vehicle having any type of real world application at all. Which you haven't.

What's your criteria here?

Because I have the same opinion as you... Except I've literally seen experimental autonomous cars driving on the street... Hence +3 years for SV, and +5 for NYC.

Taxis are not the vanguard, they're just the most visible. Autonomous trucking is underway and will likely see mass adoption before taxis.

The revolution is close - we're seeing disruption everywhere. I could empathize if this was an isolated Google vaporware experiment... but every class of vehicle has promising automation, researched by dozens of companies around the world, and implementing multiple approaches.

Most will fail, of course, but we are on the cusp of success.

This isn't a binary problem: "solved" versus "not solved". There are many parts of a solution at various states of development. Research can be replicated and reverse-engineered.

It's also unclear that why there would be one vertical-integrated company in an industry built on outsourcing. There's a lot of room for collaboration at building different parts of a car. And why wouldn't a company that builds a self-driving taxi sell it to multiple operators?

>The software and training data set an extremely high barrier to competition.

How many technological barriers to entry have survived very long?

The industry may be huge, but it's not clear that only one company will crack self-driving. I think it's far more likely every major autocompany will develop good enough versions of self driving.

Plus, these companies won't be able to hide their software behind algorithms. Someone is going to reverse engineer anything put into the open market.

Taxis are a typical market of critical mass. The bigger a company gets, the better its product inherently gets. This tends to be something that leads to monopolization. Taxi transportation was (and is) already a monopoly in many areas. And in that case they're mostly obtaining their monopoly just based on the critical mass aspect since the real benefits of an economy of scale is reduced when you have a human in the car as the price floor far higher than just the maintenance cost of the car.

And that floor is also regionally dependent. With self driving taxis there no technical reason that taxis in San Francisco couldn't cost the same as taxis in places with reasonable living expenses, at least so long as things like basic maintenance could also be reasonably automated to similarly remove humans from the picture.

And finally it will be incredibly difficult to convince consumers to swap to another brand. One taxi is, more or less, the same as another. When one works well, has an overwhelming critical mass, and offers fair prices - where is a new company supposed to compete? Also there will be minimal need to pick and choose markets, at least not in the scale you're implying. I think most of these imminent monopoly factors are going to be self evident to investors. That's going to drive absurd investment enabling companies to deploy in mass. The bottle neck will likely be how fast cars can be produced which is likely why we're seeing riding sharing come autonomous taxi companies currently partnering with vehicle manufacturers. And I've yet to touch on the final issue (though I'm sure I'm neglecting others) and that can be stated simply as what it is - mergers and acquisitions.

> And finally it will be incredibly difficult to convince consumers to swap to another brand. One taxi is, more or less, the same as another.

That's not how any of this works.

What you're saying is the opposite of obvious economic theory. It's a pretty fundamental concept that when products are commoditized people will switch to save fractions of a cent or because they're bored, or because the logo is a different color.

The only counter to this is if there are strong netowrk effects but taxis only have weak network effects for availability only.

Right, people will be driven primarily by cost. And the larger an automated taxi service is, the cheaper than offer their product - which prices that not only approach but go below the price of private ownership. And the largest players will be able to setup extensive maintenance/cleaning and other facilities ensuring that what difference does exist in quality of rides is also in their favor. This is something that companies without the same scale will be unable to compete against. And you cannot marginalize the issue of availability and response time. In a world of ubiquitous taxis, it's reasonable to expect that the average response time could begin to approach measurement in seconds. Once again that's exclusively a product of scale.

> And the larger an automated taxi service is, the cheaper than offer their product

Incorrect. As I mentioned there will be some aspects of the taxi business that will be more profitable than others. New entrants focusing only on those areas and routes with higher margins will be able to join anytime, and will. Some parts of the business will be more expensive than others (think rural vs. urban or corporate accounts vs. leisure travel) and some companies will optimize for them and thus be able to create pricing pressure.

You see this already with other transportation options that have high capital costs and availability network effects. Think airlines. The tendency in markets like this is for prices to be driven down to marginal cost.

But regardless of all this theorizing, it's much simpler to analyze, we can just look at the taxi business today. The addition of autonomy really just doesn't change the fundamental economics of the thing much at all. People cost money to hire. Machines cost money to buy and maintain. The cost structures are different but they don't have really any effect on the network effect aspect of things.

Because of the app based model and some economies of scale it's likely we'll see a bunch of national/global app based taxi hailing companies, a shift which has already for the most part happened over the last few years. Nobody in this thread has offered up any convincing argument at all for why the simple addition of autonomy really will change this dynamic at all.

From my perspective you're responding to very specific issues I've raised with hand waving. Can you give precise examples of how, again precisely, a new entrant in vehicle based taxis supposed to offer lower costs in any specific and objective way that the newer competitor would be unable to immediately match and overcome with scale?

We already have taxi services. They changed pretty sharply in recent years due to the advent of smartphone ride hailing. That was a somewhat big deal and led to some consolidation. But it still hasn't created monopolies and doesn't look likely to.

I don't have to prove anything, we already have a taxi sector and it has attributes. Burden of proof is on the people claiming that self-driving will create a winner take all market when there's neither any evidence nor a plausible mechanism by which such a thing would happen.

If we replaced airline pilots with robots would that change the basic premise of the airline business? What about bus drivers? Why? OK now why taxis?

All the things you're talking about like economies of scale and similar are applicable to taxis with people driving them. What about replacing drivers with a machine will change the network effects of the taxi business?

Answer: nothing. The two concepts are orthogonal.

The premise that self-driving inevitably leads to an all powerful global taxi monopoly is magical thinking, and I really am mystified that it has such traction here.

Do you think that your post is a reasonable to a simple request asking for any sort of logic to support your assertions? So far as I can tell, your argument seems to be a poorly elaborated upon appeal to taxis.

Even giving you the benefit of the doubt there, I don't see how this makes sense. Removing the driver is a revolutionary shift in how the entire industry operates. It also completely shifts the margins for competition, and innovation. The vast majority of cost in a taxi is the human in it. That means that in order for taxi companies the most effective means of competition on fare pricing is seeing who can pay their drivers less.

Shaving 10% off your operating cost due to scale is irrelevant when the human in the seat is 95% of your cost. Taxis in most areas are monopolized (making your choice of them as a foundation of your view even more perplexing), but Las Vegas is an exception there. There are a large amount of competitors and the net result of that was that taxi drivers just finally won a years long court case demanding they be paid at least minimum wage. That's quite silly, and sad. And something that a decade from now will be a relic of the past.

I don't care about taxis that much. But I do have a degree in economics and actually wrote a thesis, in part, on technological lock-in and naturally occurring monopolies.

Most things don't become monopolies, so the burden of proof goes the other way. And there's nothing about self-driving taxis that makes them a monopoly. For the most part a natural technological lock-in has an insurmountable network effect, the canonical example is the design of the typewriter keyboard.

In all contexts what happens is an environment where your decision to use a product makes it more valuable for all other users. Like a fax machine. Or Facebook.

Taxis are literally the opposite of this. The more people using a taxi service the longer I have to wait for a cab. Yes, you will quickly point out that the bigger the service the more that service be available wherever I tend to go.

But those two forces are in direct contradiction to each other and as such the network effects of taxi services are weak at best.

Sure you can create a hypothetical world monopoly autonomous taxi company. I can always create another company, buy a dozen self-driving taxis, and post them outside giant sporting events or clubs at closing time with graphics on the side saying download this app to rent me, and someone will do it. There goes your monopoly.

The whole freaking point of monopolies is that they happen when the switching costs are extremely high or when the supply is constrained and someone corners it. Switching costs will always be zero, and there's absolutely no chance whatsoever that one single company will permanently be the only one that creates a working autonomous vehicle.

So as neither of those conditions apply, there's no obvious force towards a permanent monopoly. In fact the reason for the government created monopolies we do see is the fact that the market tends toward commoditization and a downward spiral in prices.

The market for taxi services is unusually competitive. That's why we got to where we are in the first place.

As you seem to have acknowledge, it is untrue that the more people using a taxi service, the longer you have to wait for a cab for obvious reasons - namely cab rides end. Imagine in a city there is a cab service which services about 1 million people at a time. And there is a service which services about 100 people at a time. How likely is it that a cab is going to be concluding a ride near to your pick up point with both services? Which is more likely to have an idle cab near you? How long would you expect to have to wait for each service?

Autonomous taxis also completely revolutionize the possibilities for scale. Cars mostly only deteriorate when in active use. Massively oversupply will likely be a thing in some areas opening up possibilities like your sports event thing. People aren't going to be going to the streets to try to find a ride. There will be hundreds outside the stadium - and the cool thing is anybody using the monopolized company should be able to hop into any cab and it'll auto sync to their destination and they'll be off, rapidly replaced by another cab from the monopolist ensuring 0 wait for other users.

And then we get into cost, something you continue to inexplicably ignore. Right now there is no room for economy of scale to affect the price of taxis. I'm not going to repeat myself as I've gone into this above. Suffice to say, the biggest player is going to be the cheapest and by a wide margin.

So your dozen taxis are going to be less reliable, more expensive, likely less well maintained, require people to install yet another app, and also face a trust issue. Self driving technology will likely be closed source and proprietary for some time. So on top of all of the above issues you'll require users jump into an unknown car, from an unknown brand, using who knows what software and security standards.

And this isn't getting into the tertiary developments of self driving vehicles. We'll likely see the introduction of services (that hopefully can be disabled) that tie in the taxi to commercial agreements from the monopolist. For instance, you're passing by a coffee shop and the car offers to order a coffee for you for 50% off. You needn't pay anything (all charged through the monopolist's finance infrastructure) and a server will bring it out for you. Wait time - approximately 30 seconds. I'd assume things like that can be disabled, but on the other hand they'll be kind of nice when you're in the mood - you could also perhaps simply ask the car if there are any deals nearby on [---].

Ultimately self driving taxis will be a paradigm shift. Comparing them to current taxis (or even ride sharing services) will likely be as quaint a notion as considering there was a time when we classified automobiles as "horseless carriages."

In addition to all the profitability arguments (and I largely agree with CPLX here), we're talking about a market into which VCs have literally dumped billions recently, with little sign of abating, so any company with sustainable pricing will have to compete against companies with venture subsidized pricing.

Think this will be a winner take all. So would agree on a monopoly and Google is well positioned.