From Daimler to Didi — the car revolution

In his fantastic book, Sapiens, Yuval Noah Harari notes — “The most important things that happened in the Industrial Revolution was not the discovery of this or that particular raw material or energy source, but, the new understanding of humans that they are in fact surrounded by enormous, almost limitless quantities of energy and raw materials. The main thing we really need to do is to find and invent ways to harness and control all this energy and raw materials. What is really lacking is not energy, but ways to harness it. Over the last few centuries, every few years, or decades, scientists and engineers have managed to discover new sources of energy, new raw materials and new ways of harnessing them for our needs.”

So, our journey has human beings has involved harnessing various sources of energy and continually improving efficiency. We first heated coal and used it to power steam engines and then moved to petroleum, natural gas and used electricity to light up our world.

A key use of all this energy was to transport people and things. And, that tale began with the Paris-Rouen race for horseless carriages organized by the “Le Petit Journal.” The 102 entrants included vehicles powered by steam, petrol, electricity, compressed air and hydraulics. Only 21 qualified for the 126km (78-mile) race, which attracted huge crowds. The vehicle below was powered by an engine built by a little known company called “Daimler.”

The clear winner of the race was the internal combustion engine. As The Economist beautifully put, over the next century it would go on to power industry and change the world.

The internal combustion engine
It is important to start with the internal combustion engine because it helps us understand why the next two decades will change the transportation industry and, consequently, our lives. This was US patent #2,408,298 awarded on September 24, 1946 for “The Cummins Diesel Engine.”

The internal combustion engine is a beautiful combination of pistons, gears, rods and other parts — 2000 in number. Of course, these parts bring wear and tear with them. And, this wear and tear that has created a massive industry around servicing cars. An intriguing article by Seth Miller on Newco Shift (as per usual, all links at the end) shows why this matters.

Below is a list of the most common vehicle repairs from 2015:

• Replacing an oxygen sensor — $249
• Replacing a catalytic converter — $1,153
• Replacing ignition coil(s) and spark plug(s) — $390
• Tightening or replacing a fuel cap — $15
• Thermostat replacement — $210 Replacing ignition coil(s) — $236
• Mass air flow sensor replacement — $382
• Replacing spark plug wire(s) and spark plug(s) — $331
• Replacing evaporative emissions (EVAP) purge control valve — $168
• Replacing evaporative emissions (EVAP) purging solenoid — $184

And this list raises an interesting observation: None of these failures exist in an electric vehicle.

The electric vehicle drivetrain contains just about 20 parts. And, it doesn’t take a genius to conclude that a system with 100x fewer moving parts will be more reliable.

So, we now live in a world where a Tesla has lasted more than a million times while most cars with internal combustion engines last only 150,000 miles. This is, of course, just the beginning. Volvo has already become the first major car manufacturer to commit to phasing out all traditional engines and gas powered vehicle in 2019.

The 3 shifts
Ben Evans wrote a seminal post on how to think about cars in 2015. In it, he made 2 important points. First, he called out the fact that there are 3 shifts happening that all feed into each other.
1. Mechanical to Electric
2. Owner to Shared
3. Driven to Autonomous

Second, he explained why cars get such a large share of technology press — market size.

This chart should be taken with a grain of salt because the move to shared cars will reduce the size of this market. But, we’re nevertheless still talking about a massive market that would be susceptible to monopoly/duopoly/oligopoly via network effects in a world with large autonomous car fleets.

We started discussing why the move from mechanical to electric makes sense. Now, let’s take a leap past the move to shared cars to discuss autonomy.

Everybody wants to make a self driving car
For the longest time, the only show in town where self driving cars were concerned was Google’s self driving car project. Then, we suddenly saw an explosion in autonomous driving technology. It started with news about Uber, Lyft, and then now involves all of the traditional car companies. To give you a sense of how widespread the global self driving project is, here are a few snippets from the past few months -

• Cadillac has begun selling cars with autonomous driving capabilities
• Bosch is building a $1.1B self driving chip plant
• Jaguar is working with Lyft on autonomous driving tech — who also work with General Motors and Waymo/Google
• Cruise (acquired by GM) is already running an autonomous ride-hailing service in SF
• Lyft and Nutonomy are working on a project in Boston (Lyft’s Business Development team have been busy)
• Nutonomy and Peugeot team up in Singapore
• Intel has autonomous car facilities in Arizona and California
• Baidu wants to create the Android of autonomous driving cars
• And, we all know about Apple’s car project

This all makes sense. Given the explosion we’ve observed in artificial intelligence in the past 5 years and given moving from place to place is a key example of what AI will make cheap, this is expected.

However, we’re likely still a few decades away from getting this right. To understand why, let’s take a look at the 5 levels of automation as defined by the US Department of Transportation’s National Highway Traffic Safety Administration.

But, roads dominated with autonomous cars are a few decades away — we’re likely at the excitement phase of the hype cycle
As linear as the flow between the five levels looks, the challenge in building these systems actually looks closer to this.

These point to the challenges that lie ahead for autonomous cars. First, it is very hard to create fully autonomous cars — even in controlled environments. John Deere has been trying to build self driving tractors for the last 20 years and progress has been slow and hard.

Second, the challenge is going to be going from Level 2 systems to Level 4 systems. NASA has done decades of work involving humans overseeing Level 3 systems and the results are not promising. We just don’t have the attention span required to do this right. And, for all the comparisons with piloting airplanes, response times on the road are much quicker. Car makers get this — but this doesn’t make the task any easier.

Next, the period of transition to autonomous cars isn’t going to be easy. It would be a lot easier if we just replaced all human driven cars with autonomous cars. But, given that isn’t going to happen, we’re going to have many regulatory challenges to deal with. How safe will we require these cars to be? How will an autonomous car navigate ethical questions as it is heading for a crash because of an error from a human driven car? And so on.

Finally, as Ben Evans beautifully chronicles in another seminal post, there are far reaching second order consequences of cars. Here are a few examples of these consequences.

When autonomous cars finally become the norm, we’ll have a lot of existing systems screaming for change. It will take a while, however.

Electric vehicles, on the other hand, deserve the hype
In 2016, solar energy was available at the same price or cheaper than fossil fuels in more than 30 countries. Investment in solar naturally trended up.

You know solar has entered the mainstream when you see Ikea announcing a solar panel product.

All forecasts seem to point to 2035–2040 when we move completely away from internal combustion engines. But, in my (unqualified) opinion, we’ll see roads dominated by electric vehicles in a decade. And, that’s exciting.

Ride sharing deserves the hype too — and Didi is on a tear
I’ve barely spent time on ride sharing because their potential is obvious for those who’ve used the likes of Uber/Lyft/Ola, et al. Car owners often reluctantly buy cars in cities and these ride sharing services increase the reluctance to do.

Most of the news around ride sharing has revolved around Uber’s implosion in the past six months. Lyft, meanwhile, has used the opportunity to beef up its autonomous car partnerships. There’s a close relationship between ride sharing networks and all the experimentation around autonomous cars because of the following two step loop —

First, ride sharing networks have, well, network effects. So, once they reach a certain size, they become hard to dislodge. Case in point — Uber’s continued dominance despite all the negative press this year. This lends itself to a few location specific giants. Second, these giants will have fleets of cars that could all be converted into a massive network of self driving cars that would be easily accessible to users.

There are a collection of contenders in the ride sharing game. Uber and Lyft are the obvious ones if you’re in the west. Google is playing the game via its partnership with Lyft and with Waze’s car pooling feature. But, globally, I’m long Didi Chuxing.

In the race to dominate ride sharing, Uber raised massive amounts of funding and attempted to obliterate its rivals. Didi, on the other hand, took a different approach. It first merged with its biggest rival in China, then persuaded Uber to get out of the country (in return for a 20% stake) and has since been using its war chest of funds to invest in ride sharing networks in South East Asia (Grab), the Middle East (Careem), Europe and Africa (Taxify). Didi may be creating the ultimate network of ride sharing services.

For the longest time, ride-sharing seemed to be a battle between two opposing styles personified by Travis Kalanick of Uber and Jean Liu of Didi. While Travis focused on a testosterone fueled fight, Jean Liu used her considerable deal making acumen and an almost collaborative approach to competition to sway the pendulum decisively toward Didi.

Much to learn from her, we have.

Links for additional reading (longer than usual!)

• The Industrial Revolution (excerpt from Sapiens by Yuval Noah Harari) — on Wordpress
• The death of the internal consumption engine — on The Economist
• This is how big oil will die — on Newco Shift
• Ways to think about cars — on Benedict Evans’ blog (seminal post)
• Volvo will only make electric and hybrid cars in 2019 — on Techcrunch
• How to think about artificial intelligence — on Notes by Ada
• Everyone is working on self driving cars — Cadillac, Bosch, Jaguar, GM — Cruise, Lyft-Nutonomy in Boston, Nutonomy-Peugeot in Singapore, Intel, Apple, Baidu
• NHTSA — 5 levels of autonomy — on TechRepublic
• John Deere’s struggles with self driving tractors — on Quartz
• NASA’s experience around humans working with autonomous systems — on Scientific American
• Self driving cars are a long way away — on The Economist
• Cars and second order consequences — on Benedict Evans’ blog (seminal post #2)
• Google, Uber and Transportation-as-a-service — on Stratechery
• Solar became cheaper in 2016 — on Quartz
• Ikea introduces solar — on Ikea
• EV adoption forecast — on Bloomberg
• Waze carpooling — on Techcrunch
• Didi raises money, invests in Grab, Careem and Taxify
• Jean Liu’s approach to building a ride sharing giant- on Quartz

This is an edition of a bi-weekly technology newsletter called Notes by Ada. If you like this and would like free weekly notes via email, please just subscribe here.