Uber is working on a driverless car network that could
eventually replace all 40,000 Uber drivers in London.
In the future, the most expensive part of any transport
system is likely to be the driver, according to an analysis by
Autonomous driving systems will become so cheap that
some networks may offer them free, paid for by
Morgan Stanley believes 90% of factory jobs and 50% of
office jobs will eventually be replaced by software and
Driving is dying. If you want a job in the future you
must learn to code now.
LONDON — Since
London’s ban on Uber was proposed a few days ago, I have a
had a series of arguments with friends and strangers over whether
the ban is right or wrong. The debate has brought London alive —
on some days it feels like the only issue anyone is talking
about. Nearly 800,000 people signed a Change.org petition asking
the government to save Uber. Forty-thousand jobs with Uber are at
stake, and 3.5 million Londoners use Uber regularly.
Twenty-thousand black cab drivers would love the ban to survive
its legal challenge. Even prime minister
Theresa May weighed in.
Everyone in the capital has a stake and an opinion.
Interestingly, almost all my discussions have ended up the same
way, even when we vehemently disagree: “None of this will matter
when they start using driverless cars anyway,” one of us will
say. And we both laugh politely, the way one does when you’re
trying not think about the fact that Uber is going to drop those
40,000 drivers in favour of an army of robots.
The most expensive part of driving is the human
The company is already working on its driverless future. On
September 28 the company published a patent application for an
vehicle communication configuration system” that will allow a
central command to monitor multiple vehicles, none of which have
There is a relentless logic to it.
In a research note published by UBS on “the mass adoption of
robotaxis” last week, analyst David Lesne and his team noted that
in any transport system — private car, public transit, or Uber —
the most expensive part is the person doing the driving.
Driving to work in a private car imposes an average daily
commuting cost on the owner of €24 per day (about $24), UBS says.
In a world of robotaxis, with no need to buy a car, that cost
falls to €7.2 per day. “Getting rid of their private car would
enable the shared mobility user to travel about 10,000km per year
in a robotaxi and save €5,000 per year,” UBS calculates:
“Robotaxis will likely price-compete with mass-transit systems.
The shift towards electric autonomous vehicles, combined with
more advanced fleet optimization and servicing platforms,
next-generation traffic management and more intense competition,
should reduce the fee charged to passengers of robotaxis by as
much as 80% versus a ride-on-demand trip today. The technology to
make robotaxis a reality is already available. In this new
paradigm, owning a private car will cost almost twice as much as
using robotaxis regularly.”
That is an extraordinary thought: An Uber ride that costs £10
today — already roughly half the price of a back cab — might cost
only £2 in a few years’ time, UBS says. The cost of providing
cars without drivers might be so small that companies could offer
rides for free, UBS speculates, and make money on the advertising
That is a real problem if you’re an Uber driver.
The smart money says that in the short-term Uber will
successfully challenge the ban and reach a compromise that will
allow it to operate without interruption. The real threat to Uber
drivers comes from Uber itself, and its long-term plan to get rid
of all drivers. (Former
CEO Travis Kalanick told Business Insider this was the plan back
And it’s not just 40,000 Uber drivers.
The scale of the jobs carnage will be vast
Any job that involves a human behind a wheel will be threatened
in the next 20 years. Driverless technology is being developed by
a dozen or more large tech companies, including Google (Alphabet)
Apple in its Berlin lab.
The scale of the carnage in the jobs market will be vast. It’s
not just drivers. It is any job where artificial intelligence can
do it cheaper than a human. The research team at Morgan Stanley
sent a note to clients last week that calculated some of those
savings (i.e. job losses):
- Annual salary of a regulated financial institution IT
operations worker in New York: $70-80,000.
- Annual licence fee for a robot doing the equivalent work of
up to five humans $8-11,000.
Morgan Stanley says 90% of factory jobs and 50% of office jobs
are at risk of disappearing in Europe and the US:
“… not only the jobs with routine/repetitive tasks are at a
high risk of automation (up to 90% we estimate) but AI also puts
jobs involving cognitive skills at risk although the probability
is lower at up to 40%. We estimate that 50% of the US/European
white collar jobs (including office and clerical jobs) are at the
risk of automation. We think it is unlikely that job losses would
reach this level given that this is a long-term forecast (we
assume only c.16% penetration by 2022) and new skills and jobs
will be created over time. However, it is clear that some jobs
will be permanently lost, which should impact staffing revenue
and earnings growth.”
What is to be done?
There is one growth area for jobs that pays a lot better than
driving: Tech jobs. Specifically coding.
There is no unemployment in tech. Companies are desperate for
qualified workers, and the world can’t produce coders or
engineers fast enough. Even entry-level coders can instantly
enter the middle class. A random sample of the
first page of London tech jobs on Indeed shows that minimum
starting salaries are £45,000 ($60,000). Salaries escalate
quickly from there. Highly qualified individuals with a few years
expertise can name their price.
All these driverless cars, these fleets of autonomous vehicles —
and anything else that uses tech, i.e. everything— will
need people to write and manage software, in just the same way
that every office that once had a typing pool now only employs
people who know how to use email. In the future, being able to
code will be as important as being able to read. (It might be
more important, given that Siri and Alexa will be able
to just read stuff to you without you looking at it.)
If the Uber ban — intended to begin this weekend — teaches us
anything, it is that parents and schools need to teach their kids
how to code software. In fact, I wouldn’t wait for your kids’
school to get its act together. If I were you, I would sit them
in front of an online course as soon as you can.
Parents, teach your kids to code. Now.