Yes, Uber can – and will – get worse

It might be hard to imagine how things could get worse for Uber – but they will!

rtx2i8y7 2
Reuters Staff (Reuters)

It might be hard for some to imagine how things could get any worse for Uber. CEO Travis Kalanick was sent home for a couple of months. His deputy Emil Michael was terminated.

This is after Uber fired VP of Autonomous Vehicles Anthony Levandowski last month for not complying with a court’s directive in Google’s trade secrets lawsuit against them.  Oh, then there’s the 20 employees canned over a sexual harassment scandal.

And in the middle of an Uber all hands meeting – to report on the sexual harassment issue – Board member David Bonderman made a chauvinistic remark to Arianna Huffington, and subsequently resigned.  All of this came after Uber President Jeff Jones resigned in March, after only six months on the job.

There.  Did I get it all?  Doesn’t really matter.  It’s going to get a lot worse.

Last year, Travis Kalanick bet the company’s future on autonomous vehicles.  Firing Levandowski was important, as it will likely chase Uber’s best and brightest out the door.  Witness Sherif Marakby’s departure to his old stomping grounds at Ford when their new CEO Jim Hackett came calling his first week on the job.  Hackett replaced Mark Fields, whom the Board believed wasn’t moving fast enough into the world of electric and autonomous cars.

That’s what Uber is about to learn the hard way.

With nobody at the helm, moving quickly is no longer an option.  For Uber, the timing could not be worse.  It is already having trouble recruiting engineers as the scandals have piled up. But a far bigger problem lays ahead; one that hasn’t been addressed yet.

If Uber’s future is in autonomous vehicles, what is it going to do with all of that data?

Digital exhaust

An oft-cited statistic states autonomous cars will spin out about 4 Terabytes (TB) of data per day. That sounds like a lot.  But if you look closer you can tease out some very important assumptions about that number and what it really means.

It’s 4TB a day based on a 1.5 hour daily average commute.  Average.  So what we’re really talking about here is 2.5TB of data per operating hour, (4.0/1.5 = 2.5).  So let’s go with that number as opposed to an artificial average across the country.

Taxis, buses, and probably delivery vehicles will be the first mainstream, commercial applications of autonomous vehicle technology in the U.S.  We’re seeing some deployments now, through impressive demos under way via NuTonomy in Singapore, tractor trailers in Germany, and driverless taxis for the elderly in Japan.  These are all technology demonstrations right now – ironing out the kinks.

But, let’s do the math.  2.5TB an hour over a ‘typical’ eight-hour shift is 20TB of data per shift.  But ‘shift’ is a human-driver centered concept.  These vehicles are intended to be autonomous.  They might begin with a driver behind the wheel for this early period, but eventually they’ll zip around on their own.  So let’s consider a vehicle that switches out drivers every eight hours – that’s three shifts per day.  2.5TB times a twenty-four hour day is 60 terabytes!

That’s every vehicle, every day.  A popular physical measurement for data storage is one terabyte equals an eight foot high tower of CD’s!  It’s also cutting down 50,000 trees to create 250 million sheets of paper (printed on both sides), a stack that would reach 10 miles high!

So, given that Travis is going to be home on his Xbox for the next few months, who is going to recruit the data scientists required to analyze 60TB of daily data for each taxi and delivery vehicle?  Uber didn’t buy the self-driving truck firm Otto, the very heart of that Google lawsuit, for no reason.  Travis sees a lucrative future in automated delivery and service vehicles.

But what data scientist in their right mind would join a company so mired in controversy, so utterly devoid of adult supervision, when there are hundreds – thousands – of admirable companies out there looking for finger-flinging data magicians to do that work?  

Yes, Uber has a huge market capitalization – higher than Fords, as widely reported.  But so what?  How does that pay a newly graduated data scientist’s bills?  How does that answer his thirteen year old daughter who asks ‘Why do you work there Daddy’

There are simply too many other good opportunities out there where people in high-demand roles don’t have to be embarrassed or feel like they must explain themselves, to take a job.  This is where Uber is about to find themselves.

Sometimes money isn’t enough

Whomever takes over for Travis Kalanick, whether it is temporary or permanent, is going to have to address this.  Yes, there are a multitude of recently vacated positions that must be filled.  There will likely be more.

But a solid data science team will be among the most critical hires Uber can make.  Without dedicated, engaged, and free-to-ask-hard-questions data scientists, that outrageous market cap is going to quickly fade away.

Uber’s board will be approached about selling the company outright.  Ford comes to mind immediately.  GM too.  Sure, there could be a tech giant suitor, perhaps even one of the new Chinese titans like Baidu or Tencent.  The issue may be less about wanting Uber itself and more about preventing a competitor from acquiring it.  Uber doubled its revenues last year to $20BIL, according to a Bloomberg report, but is still ‘bleeding cash’.

That blood is in the water.  The only question now is, how fast will the sharks show up?  Don’t expect good data scientists to willingly jump into that feeding frenzy!

This article is published as part of the IDG Contributor Network. Want to Join?

New! Download the State of Cybercrime 2017 report