They promised us driverless cars. What happened?: Don Pittis


The name of the Scottish indie band We Were Promised Jetpacks alludes to disappointment that the future isn’t all it was cracked up to be.

Well, you can now buy a functioning jet pack if you have a spare $600,000.

A self-driving taxi to pick you up at your Canadian home? That’s still in the future. Some experts say the distant future.

That’s certainly not what it felt like a few years ago, and enthusiasts who expected things to move much faster are beginning to sound discouraged.

When I wrote about robot taxis in 2015, I quoted experts from Google’s self-driving car project — since spun off as a subsidiary under the name Waymo and still considered by people I spoke to as the one to beat — saying autonomous vehicles would be “available to the public in two to five years.

Now, five years later, Waymo driverless taxis do operate on the streets of Phoenix, Ariz., in a huge (130-square-kilometre) carefully monitored zone, keeping vehicles inside the designated area, limiting who can use them, keeping track of user reactions and watching for potential errors.

“Members of our early rider program are now being matched with these fully driverless cars when they hail a Waymo vehicle using our app,” Waymo boasted in an email in response to my query, though the program was interrupted by the COVID-19 pandemic.

What’s the delay?

When the company was asked: “What do you see as the current timeline to get Waymo-directed vehicles widely released for autonomous operation in the northern U.S. and Canada?”

It wouldn’t specify a timeline and instead said: “We’ll continue to learn, advance this offering, and make it available to more and more riders across both our early rider program and public service over time… From our decade of experience, we’ve learned that this technology will come to the world step by step, not in a flip the switch moment or at a definitive time.

“We think our gradual approach to introducing self-driving technology to the world is the safest and most responsible way to go.”

There are good reasons why the company might be reluctant to specify. Along with causing disappointment because the future may not be coming as quickly as we had hoped, the answer to that question has a direct bearing on the company’s market valuation.

Only two years ago, financial giant Morgan Stanley declared Waymo was worth a notional $185 billion US, but a year later as the prize seemed to slip away, analysts cut that valuation to $105 billion.

A subsidiary of Alphabet (formerly known as Google), Waymo is widely considered the front-runner in driverless car technology, and the Canada Pension Plan has a stake. (Waymo)

This past spring, when Waymo CEO John Krafcik announced a $2.25 billion outside investment in the business, including participation from the Canada Pension Plan, it effectively valued the company at $30 billion. On Monday, the company said that investment ended up amounting to $3.2 billion.

CPP Investments sees its stake as a long-term investment, which may be a good thing.

That said, many companies, university research departments and governments continue to pour money into the driverless concept.

Just last week, Volkswagen took a stake in TuSimple, a U.S. company considered a leader in driverless trucking. The firm promises completely driverless capability by next year, and it is still hoping to begin building a driverless network on specialized routes by 2024.

Raquel Urtasun, a high-flyer and co-founder of Canada’s Vector Institute for Artificial Intelligence, lists herself as head of Uber Advanced Technologies Group in Toronto, but neither she nor Uber were able to answer questions by deadline about why autonomous road service has been delayed.

Disruptive but disappointing

Driverless vehicle enthusiast and engineer Paul Godsmark told me three years ago that the arrival of driverless trucks was imminent. He still thinks they will be disruptive when they finally get here and said he is not the only one to have been disappointed that things haven’t moved faster.

Godsmark points out that for Uber, being a leader in self-driving technology was seen as critical to its future. As the company continues to struggle with its expensive human drivers — unlike other tech companies that have been soaring — Uber’s shares trade below their early highs.

“I think it is still an existential threat to Uber,” said Godsmark, an engineer with the Canadian autonomous vehicle consultancy CAVCOE.

Godsmark sees driverless cars following the path of Gartner’s Hype Cycle, where a frenzy of expectations is followed by disillusionment — in this case partly because of the problem of human safety.

According to Gartner’s Hype Cycle, we may be in the trough of disillusionment after inflated expectations over driverless technology. (Gartner Inc.)

“You could create a vehicle that was 99 per cent safe eight, nine years ago, but no one wants to go in a vehicle that would kill you every hundred rides,” he said.

Automotive consultant and engineer Matthew Klippenstein described it to me as the tyranny of nines, as engineers increase safety levels from 99.9 per cent to 99.99 and so on, unsure when it is safe enough.

Mary Wells, dean of engineering at the University of Waterloo, is one of those who thinks robot-driven cars and trucks mixed with normal traffic on Canadian roads remain a long way off. When it comes to basic engineering in self-driving vehicles — where, she said, Waterloo is a global leader — proponents often say we are already there.

“But they haven’t looked at it with the lens of the other complicating factors,” Wells said.

Among the complications, she lists things like computer security, Canadian winter weather, the legal and insurance regime, artificial intelligence ethics and the astounding unpredictability of human behaviour — such as the recent case of a Tesla travelling more than 140 km/h on an Alberta highway while the driver had a nap.

Even more basic, Wells said, is the question of how much responsibility we want to wrench out of the hands of humans and turn over to AI.

Wells sees enormous value in the quest for autonomous vehicles. The technology is creating many spinoffs and is already widely used in specialized environments, such as warehouses, mines and automated trains that don’t have to contend with quirky human drivers.

Certainly one technique, Wells said, might be to designate some areas as autonomous-only, where the vehicles all follow the same rules without the presence of pesky humans.

She sees the technology expanding into places like agriculture, driverless shuttles and autonomous-exclusive truck lanes — and only much later going into everyday use on Canadian public roads. And as for the high expectation still being expressed by some proponents, the Waterloo engineer remains dubious.

“In my lifetime, are we going to see widespread use of autonomous vehicles? I’m not sure.”

Follow Don Pittis on Twitter: @don_pittis



Leave a Reply

Your email address will not be published. Required fields are marked *