Google is poised for a comeback in robotics, and could take on Amazon

Boston Dynamics’ Spot, a
robotic dog that’s been the star of many viral YouTube
videos.

YouTube

• In 2013, Google attempted to build a groundbreaking
  robotics division, but the effort was scrapped not long after
  it began.
• Despite that earlier stumble, Google is bolstering its
  ranks of roboticists at X and Google Brain.
• Some have speculated Google may be building a rival to
  the “domestic robot” that Amazon is reportedly working
  on. 

---

Robots may be making a comeback
at Google.

Google made a big bet on robotics
starting in 2013, when it acquired nine of the sector’s top firms
in an effort to stake out turf in a burgeoning sector — much the
same way that the company is investing now in artificial
intelligence. After just two years, however,
Google scrapped the project, known internally as Replicant.
The thinking in the industry was that while Google continued to
operate several robotics divisions, any grander ambitions were
dead.

But the signs now are that the
robotics units at Google Brain and X — the search giant’s
artificial intelligence unit, and the “moonshot factory” formerly
known as Google X, respectively — are much revived. It’s not
clear what, exactly, this revamped Google robotics project could
look like, but one source tells us the robotics-industry gossip
is that Google is working on a rival to the “domestic robot” that
Amazon is
reportedly building. 

In June, Google Brain brought
back Ryan Hickman, who helped lead some of Google’s earliest
robotics efforts, going back to 2010. According to Hickman’s LinkedIn
page, he’s now in charge of product and operations “on a new
robotics effort.” And as far back as December, X hired Liz Murphy,
formerly an autonomous-robotics expert at Apple. Hickman and
Murphy did not respond to questions from Business Insider about
this story.

Meanwhile, in recent months, X,
the experimental hardware lab and Google’s sister company under
Alphabet Inc., has
posted a spate of new
robotics-related positions on its
site.  
Most of
the ads start this way: “We believe there are many problems in
the world in which robotics could play a significant role in
making it easier, faster and safer for people to get things
done.”

A source at a respected
university told Business Insider that Google has been recruiting
some of the nation’s most talented robotics students.

Brian
Ach/Getty

In a statement to Business
Insider, a spokesperson for X confirms that the company has been
hiring roboticists and AI experts. 

“X is working on a number of
moonshots with robotics at their core. We’re hiring talented
engineers with a background in robotics and ML who are interested
in exploring how robotics combined with machine learning can help
solve some of humanity’s biggest problems,” says the
spokesperson. 

That Google might be willing to
try its hand again at robotics isn’t totally a surprise to
industry observers.

“No, I wouldn’t be surprised if
they’re going at it again,” said Alex Broadbent, a former program
manager at Boston Dynamics, one of the robotics firms Google
acquired in 2013. “They have new leadership and they probably
want to go at it differently. They’re probably asking themselves
how do we salvage this. We got all this brain power and
technology, so how do we put it all together and change the
world? Again.”

Google’s quest to master robotics

For a brief period, Google
appeared poised to breath life into robots and transform all of
our wildest sci-fi fantasies into reality.  

In 2013, Andy Rubin, the Android
cofounder and a prized executive at Google, oversaw the
acquisition of nine top robotics firms. Among them was Boston
Dynamics, the company famous for posting jaw-dropping videos to
YouTube of automatons walking through a snow-covered woods,
picking themselves off the ground, or loading boxes. What Google
hoped to build was never revealed but
emails showed Rubin hoped to launch a product by 2020.

Then, less than a year later,
Rubin left Google to start a hardware incubator called
Playground. From then on, the robotics project
appeared rudderless. In 2016, some of the companies Rubin
acquired were sold,
including Boston Dynamics, which has been owned by SoftBank
since 2017. Others were folded into X.

There’s a case to be made that
Google made some serious missteps that hurt not only the company,
but the entire field of robotics. In October,
Bloomberg published a story indelicately titled “Google Has
Made a Mess of Robotics.”

In April,
Bloomberg reported
that Amazon was making progress on a robot
codenamed “Vesta,” and that unnamed sources told the news outlet
might be a “sort of mobile Alexa, accompanying customers in parts
of their home where they don’t have Echo devices.”

If Google is playing defense and
creating a mobile version of its Google Home speakers, it would
make sense. Amazon’s Echo and Google Home are battling to become
the dominant home smart speakers.

Robots for the home are hard to
build

In this scenario, hiring Hickman
also makes sense. After he departed Google in 2016, he eventually
cofounded TickTock, an AI and robotics startup that tried to
crack the onerous problem of creating robots that can navigate
home environments. TickTock didn’t get far. After a year, the
company closed down after attempting four different robot
concepts, according to

Wired.

Starting in April, the day after
Bloomberg published the story about Amazon’s Vesta robot, Hickman
began posting essays to Medium dealing with indoor robots. The

first was titled “Why an Echo Show on Wheels makes sense for
Amazon.”

In June, presumably shortly
before Google hired Hickman back, he wrote about the complexities
involved with supplying

robots with better
vision

.  

“I want a fast-moving home robot
that doesn’t crash through my glass door,” Hickman wrote. “I’d
like to walk into a room and not trip over a robot that failed to
get out of the way. I want to use all the verbal queues I do with
a real dog (e.g. “sit”, “stay”, “heel”, “lay down”, “come here”)
to command my robot on where to be. This requires a level of
spatial-reasoning AI that doesn’t exist today.”

Hickman indicated that he’s
optimistic the challenges can be overcome. Others are far less
sure. Almost everyone interviewed agreed said the sensors needed
for a robot to navigate a home are too expensive and clunky.
Others said they doubted that Google, after its previous botched
attempt, is interested in any in-home robot.

 “Last time Google hired
like crazy (in robotics),” said one former employee, “it
didn’t work very well for them.”