Over the course of a year, a group of research scientists at Google’s secret lab, X Development, were obsessed with flies’ brains.
The team believed it could build more efficient neural networks — computer systems for processing information — by modeling the brain structure of flies and other organisms that had evolved over millennia to perform specific tasks.
The project, known internally by its codename, “Valkyrie,” was the typical science-fiction-sounding idea that Google’s self-described “moonshot factory” was designed to dream up. The potential for Valkyrie also seemed vast, from controlling drones to cleaning up satellite imagery faster than existing algorithms could.
But last spring, X Development leaders shut down the project because it lacked a clear path to monetization.
Google designed X Development, or X for short, to make technological breakthroughs and work on humanity’s most stubborn problems, from transportation efficiency to the climate crisis. Granted, not every solution pans out, and killing projects is often celebrated. But amid a lack of successful launches and a tech industry facing economic headwinds, insiders say the lab is now hitting the hard-reset button and rethinking its purpose within Alphabet, Google’s parent company.
Insider spoke with a dozen current and former employees and reviewed internal documents that show the company is moving further away from its original premise, scaling back on its most ambitious projects in search of bets with more immediate revenue-generating potential.
These sources spoke on the condition of anonymity because they’re not authorized to speak to the press. Their identities are known to Insider.
Google declined to make the leader of X, Eric “Astro” Teller, available for an interview. An X representative declined to comment for this story.
Formed in 2010 as Google X, the unit was the founders Sergey Brin and Larry Page’s personal sandbox for radical creativity, cut from the same cloth of once renowned research labs like Xerox’s PARC and Nokia Bell Labs.
The earliest employees at X were tinkerers encouraged to work on far-out ideas like space elevators, levitation, and even teleportation.
Even as the division separated from Google under the Alphabet structure in 2015, the unit thrived on being scrappier than its mothership. X swapped standard corporate titles like “Communications Lead” and “Manager” for grander and more-nebulous names like “Factory Whisperer” and “Firestarter.” Teller, the division’s leader, isn’t a CEO, but a “Moonshot Captain.”
Inside the division’s offices in a former mall in Mountain View, California, self-driving cars whiz around employees eating lunch and robots whirr around desks.
Teller, who oversees a division of about 500 full-time employees, is known to roller-blade from meeting to meeting as he waxes poetic about blue-sky, risk-taking, and moonshot idealism.
Unmoored from the worries of profitability and Google’s core search-and-advertising business, X tells employees that the “next Google” will be born in its halls.
“It fit into the worldview that Larry and Sergey projected: Just leave us alone and let us innovate, and you won’t be sorry,” a former employee said.
Now Alphabet is under increased pressure to show value in its “other bets” segments — the constellation of non-Google companies that sit under the Alphabet umbrella, such as the life-sciences unit Verily — that continue to incur billions of dollars in operating losses each year.
But over its first decade of existence, X has struggled to produce big winners. Its internet-balloon unit, Google Loon, meant to bring people in rural areas online, was shuttered in January 2021. Its drone project, Wing, while still active, has been slow to gain traction. Waymo, Alphabet’s self-driving-car initiative, is starting to generate revenue, but a profitable business model could be decades away.
The realities are starting to butt up against X’s ambitions. By 2030, Teller wants X to deliver an internal rate of return greater than 26% for Alphabet, according to an internal manifesto written by Teller and seen by Insider.
He has also told employees he wants to spin out 20 companies by 2030, with at least four being worth more than 5% of Alphabet’s market capitalization, according to the same document.
In November, Bloomberg reported that in pursuit of these new quantitative goals the division had struck up a partnership with the farming company Driscoll’s to use Google’s artificial intelligence and robots to try to create tastier strawberries. The deal reflects more practical ambitions than the fantastical science-fiction projects the division was designed to dream up. Selling machine-learning algorithms to farming companies is far more profitable in the near term than spending resources on taking humans to Mars, an idea that X employees have occasionally toyed with.
“They’re no longer moonshots. They’re Ruth-shots,” one former senior employee said, referring to Alphabet’s chief financial officer, Ruth Porat, who has curtailed much of Alphabet’s lavish spending, including research-and-development projects, since joining Alphabet in 2015.
X also lost its biggest cheerleader when in late 2019, Brin — along with Page — stepped back from Alphabet.
“We lost our guardian angel,” another former senior employee said.
These changes have also led to the unit becoming slightly less freewheeling as an uptick of venture capitalists and management consultants join to help accelerate the projects’ commercialization pace.
“You’ve seen a rise of traditional, much more business-oriented roles,” one former employee said. “While they’re not driving the org, they now have a much larger say.” While X used to be a haven for engineers and doctoral residents, insiders said the number of business analysts, program managers, and directors had increased in recent years.
This could help projects reach profitability much faster, but some employees worry it’s killing potentially world-changing ideas before they can flourish. Two said that X had begun to kill projects earlier in the pipeline than it used to.
“Bringing in VCs who didn’t understand what it means to be free with technology has driven out a lot of the folks who joined for the moonshot thinking,” one former employee said.
“Moonshots are not VC-friendly,” another said.
Some insiders, however, argue that as Alphabet increasingly curtails its spending, this type of commercial-first thinking is necessary for X to survive.
“I think the infusion of VC people is a good thing,” one former senior X employee said. “Most of these folks are deep tech. Where are the businesspeople to test viability in the marketplace?”
Speaking at an MIT Technology Review event in November, X’s chief technology officer, Benoit Schillings, said X used to explore projects “without much rigor on, ‘Would this really work?'” He said X had since become “more and more rigorous” on producing moonshots.
Perhaps X’s biggest inflection point, insiders said, was Brin stepping back from Alphabet.
Many of X’s initial ideas came from Brin and Page. Brin would roam the halls in Crocs and Lululemon pants, checking on his latest toys. “He has super high energy,” one former employee recalled. “He’d come in and rabbit-hole on one technical thing.”
Previously, Brin would protect projects he was close to, including Makani, an idea to extract energy from wind kites. A former Makani employee recalled that Brin continued funding the project even when it failed milestones he set. X shut down Makani in 2020 after Brin stepped back from Alphabet.
“Sergey might have put his finger on the scale of the projects every once in a while. Once he left, that finger disappeared,” a former employee said.
“It’s no longer Sergey’s playhouse,” another X-er said. “It’s an org that exists to make money — and it’s trying to use Google’s competitive advantages to make money.”
While X was previously known for its wild hardware projects like wind-energy kites and the notorious public failure that was Google Glass, insiders said the recent shift toward AI-friendly projects like the Driscoll’s partnership are also a reflection of Alphabet CEO Sundar Pichai’s vision for Google. One of X’s recent “graduates” is a project that teaches artificial intelligence to write and fix code, which was moved into Google last year.
A focus on more commercially-friendly ideas may put some of the division’s biggest projects in jeopardy, including A-Life, a secret years-long effort to engineer synthetic life through the study of cells. Hardware projects are now getting less attention too, insiders say. Wolverine, a device to give humans augmented hearing, left X and Alphabet last year and is raising money outside as a startup. Meanwhile, projects focused on climate are still getting some support.
“Astro is under the gun to produce some winners, and produce them quickly,” one employee said.
But the bet on AI might be paying off. The team behind a strawberry-counting project, Mineral, which uses machine learning and camera-laden robots to improve farming, has held discussions about graduating into an “other bet” of Alphabet as soon as this year, according to a person close to those talks.
Some employees speculate whether X might last as a stand-alone unit. And as Alphabet’s leadership makes more pragmatic bets amid a souring economy, employees across the company are bracing for layoffs. Some are wondering whether X will keep its headcount over the coming years.
“I don’t think X will last,” a former Alphabet executive who was part of Page’s L team of top leaders said. “X has not produced. There’s no product from X that you use today. You can’t spend that much money and not produce anything and say it’s a success.”
In September, the company cut about half the projects in its Area 120 division, another research-and-development unit within Google that spins up new product ideas. Only a handful of AI-first projects would remain, keeping in step with Pichai’s broader vision for the company.
TCI, a Google activist investor, recently called on the company to cut back on its spending on “other bet” projects. Still, by most measures, Google’s business is still a smash success — it raked in $69.1 billion in revenue in its third quarter despite a slump from market headwinds.
Most likely, as part of Teller’s move away from X’s flashier projects, it will lean into safer ideas that are less science fiction and more a reflection of Google today. They just might not shoot for the moon.
Epitomizing this change in direction, one employee told Insider there used to be a joke at X that to get leaders to approve a project, you made it a robot or made it fly. Now, the person said, “the new joke is, you take some existing problem in the world and sprinkle some AI on it, and it’s an X project.”
Are you an X or other Alphabet employee with a tip? Contact Hugh Langley via encrypted messaging app Signal at +1 628-228-1836 or email at hlangley@insider.com.
Read next
11 January, 2023
0 Comment
158 Views
Google's Moonshot Factory Is Coming Down to Earth – Business Insider
by Jason Peters
Over the course of a year, a group of research scientists at Google’s secret lab, X Development, were obsessed with flies’ brains.The team believed it could build more efficient neural networks — computer systems for processing information — by modeling the brain structure of flies and other organisms that had evolved over millennia to perform... Read More