Robots and the real world

Image: https://www.youtube.com/watch?v=YCY95dnDbn0

Image: https://www.youtube.com/watch?v=YCY95dnDbn0

All over the world, ageing populations and changing attitudes are making it harder to hold on to workers, believes Dr Rodney Brooks. Brent Balinski spoke to Brooks, founder of Rethink Robotics, about some demographic and technology shifts to watch out for and why it’s best to be realistic.

In the markets where Rethink sells its collaborative robots, ageing workforces and a shortage of younger workers available to replace retirees are a combination troubling manufacturers, according to founder, chairman and CTO, Dr Rodney Brooks.

It’s a problem that’s been brewing for many years, and an issue in China which he picked up on early in the millennium as co-founder of iRobot (which still produces its Roombas there, at a rate of around two million annually). The same goes for Adelaide-born Brooks’s adopted home of the US, he said. Ditto for Europe.

“They just don’t know what’s going to happen in some of those factories in the midwest states; the workforce is ageing out,” Brooks told Manufacturers’ Monthly recently.

“In the US and Europe the average age of factory workers is going up all the time. BMW has special programs to retain the older workers. They’re not trying to get rid of them – they’re trying to keep their elderly workers! They’re the ones with all the knowledge.”

In much of what he’s seen, people simply don’t want factory floor jobs as much as they used to.

As a roboticist, his solution is unsurprising: more technology, specifically more robots. Industry 4.0 is topic some get carried away with, but not Brooks. Most of the world’s manufacturing is low tech, he said, taking place in brownfields factories with low-skilled, high-turnover workforces. He dismisses the Industry 4.0 vision as “top down” and based on some often unrealistic assumptions.

The factory of the future might have a seamless flow of information across value chains and self-optimising machinery, but the factory of the present – in China and elsewhere – can be cramped, run by ageing equipment, and not even networked. Creaky, decades-old machines will be kept as long as they still do the job.

He recalled the GE Minds + Machines event where he made the point as a presenter: “Some GE [General Electric] Digital people came up to me and whispered, ‘We’ve still got factories like that, too.’ But it doesn’t fit the narrative of stuff coming from above and all the bits whizzing everywhere.

“Investment has to provide incremental value. ‘Invest a whole lot and then you’ll get some undefined value’ [is unappealing].”

Three shifts to watch

According to Brooks, there are three big disruptions to expect in the industrial world.

One is the “demographic inversion” and accompanying labour issues mentioned above.

Another is high numbers of 3D sensors, with high quality and low prices driven by developments in the video game industry (Microsoft Kinect is the go-to example).

Then there will be a generational shift in how industrial equipment is designed, incorporating regular software updates. Brooks expects this concept – comparisons include Tesla’s car and Apple’s iPhone software updates – to become common on the machinery on plant floors, and to provide tangible benefits to users.

“Traditional equipment… had embedded software and it’s buried in its ROM or something, but these devices live through regular software updates,” explained Brooks, pointing to his phone.

Image: www.australiaunlimited.com
Brooks and Sawyer (Image: www.australiaunlimited.com)

“So you buy the piece of hardware, but its capabilities continue to evolve over time… Imagine the results of new deep learning algorithms, trained on millions upon millions of things elsewhere, downloading that onto the machine and now the machine can deal with seeing new parts or new things that weren’t invented at the time the machine was built.”

Baxter and beyond

According to Brooks, Rethink – which last week gained an extra $US 18 million in venture funding for its product development and expansion plans – saw a “massive increase” in sales in 2016 driven by its Sawyer robot. Sawyer was the company’s first for the world’s manufacturers (its predecessor, Baxter, was only designed for US factories).

The last time this magazine spoke to him, in 2015, they had clocked up a claimed 1,000-plus sales. This number is currently in the “thousands” said Brooks, whose company announced several new distribution deals over the last year.

The topic of technology and its impact on jobs – in manufacturing and elsewhere – continues to be a thorny one. The realists will concede that manufacturing will remain important, but employ fewer and fewer people in total as automation drives productivity upwards.

Within the United States, president-elect Donald Trump and his supporters are contending that it’s possible to bring lost jobs back. Recent high-profile cases including Carrier, Ford and General Motors have emboldened those sharing the view, though opinion is divided on how productive tariff threats are for the country’s manufacturers.

(For context’s sake, this interview took place on December 5. Brooks was in Australia at the time for UTS Business School’s Future of Manufacturing Forum.)

The roboticist – who started his current company with a vision of making American manufacturing more competitive – said he understands the optimism around bringing production and jobs home. However, he feels it is misguided.

“I’ve seen press stories – some woman in Toledo, Ohio saying ‘it might take Mr Trump two months to get steel production back’,” he observed.

“It’s just totally unrealistic. And I don’t know whether it was thought out at the high level any better than that.”

In manufacturing, as with everywhere else, nostalgia has its place, but it’s always best to be realistic.

“I had the same thing when I was running computer science at [Massachusetts Institute of Technology]. A lot of people, older faculty members, were like ‘why can’t it be like in the 60s when DARPA [Defense Advanced Research Projects Agency] gave us a block grant and we decided what we wanted to do,” said Brooks, who headed  the MIT Artificial Intelligence Laboratory and then the Computer Science & Artificial Intelligence Laboratory from 1997 until 2007.

“Can you make that happen, Rodney? Can you go to Washington and make that happen?’ I can’t make that happen. That’s where the techno-political situation was in the 60s. It’s a different world now and no matter how much you liked it then, it isn’t coming back.

“We have to figure out how we operate in the new world. I’m not trying to denigrate any particular group of people; the MIT faculty had exactly the same nostalgia: ‘Why can’t it be like it was back then?’”