Home > News content

Uncover Google's "moon project" factory X: build the next Google

via:网易科技     time:2018/7/13 0:36:59     readed:51

At 6:40 in the morning, a car horn went off three times. "Add air!" A man in a helmet and a fluorescent vest shouted. There was a hissing sound, and the helium began to flow. A large container like wood is piled from a nearby truck, and helium passes through a series of hose up to 55 feet high, then through a copper tube, into a plastic tube that falls down to the ground.

This is a Wednesday in late June, in a remote mining town across the I-80 highway in Winnemucca, Nevada. Winnemucca is the testing base for the Project Loon project. The grand project was launched in 2011, with the aim of bringing the network to the sparse population of the earth, because of the complex terrain that is difficult to build a traditional network base station. Loon plans to fly the balloon up to 60 thousand feet above the ground, instead of building and maintaining ground facilities, and each balloon transfers the network to nearly 2000 square miles of area.

Winnemucca is located in the north of Nevada, with a small population. This place is very suitable for exploring how to make giant, transport network balloons fly to 60 thousand feet in the air.

Loon is testing in Winnemucca because the sky is very empty, and there is an airport, and the top people can fly directly from Palo Artaud on private planes, not far away. Today, the team is testing a new version of its communication system, which may be able to support users 10 times its current settings.

After half an hour, the balloon can be released. It is now fixed by a red crossbar, with three walls blocking the wind. Under the command of an engineer who waved yellow remote control device, the structure of the Big Bird was rotated 90 degrees to the left. Like the lion king, when the opening of the lion king raised the newly born Simba, all the arms of the crane pushed up the balloon upward. When the balloon carries the weight of the load - solar panels, antennas and all kinds of electronic equipment - it pause for a moment. Then it floats in the wind, climbing 1000 feet per minute. It's quite spectacular.

Loon and Wing are both graduating

When Nick Kohli joined Project Loon in 2012, he was running around looking for and collecting from the Mojave Desert. Falling balloons from rural Brazil and the coast of New Zealand. Loon used to be part of search company Google's unit X, which aims to foster bold projects to apply emerging technologies in innovative ways to very difficult issues. One is self-driving cars. (Google changed its name to Xerox when Google restructured its parent company, Alphabet, in 2015.)

Collie, not a typical Google employee, has the qualification to survive in a catastrophe. He did not get into the medical school, so he accepted the training of the emergency room technician - a background, with his pilot's license and the eight year search and rescue experience in the Sierra Nevada, to make him the kind of talent that Loon needed. This practical combination of skills and operational vision makes him one of the many new talents that X needs to fulfill its mission. The mission of X is to extend the Alphabet's tentacles beyond the mobile phones in your laptop and pocket.

With the help and resources of Alphabet, Cole (now in charge of flight operations) has been able to witness the entire evolutionary process of Loon: from watching a balloon flying hundreds of miles from the first, to such a launch as today has become commonplace. This is just another step towards a complex system that envisages Loon.

Today, X announces that Loon wants to "graduate" - becoming an independent company under Alphabet, marking an important step towards achieving this mission. Like another project, Wing, which uses unmanned aerial vehicles to deliver goods, Loon will also set up an employee team and set up its own human resources and public relations team. Its leader will receive the title of chief executive whose employees will get an undisclosed share of the company's success. Creating income and profits will be as important as changing the world.

Loon and Wing were not the first ones to get a diploma from X. Verily, a life science project that monitors blood sugar levels with contact lenses, has made this leap back in 2015. In December 2016, the autopilot project also made a leap, named Waymo. Chronicle, a network security project, was promoted to an independent company in January this year.

Loon and Wing are both graduating - two ambitious projects - a watershed in X, perhaps marking the time for the mysterious research and design department to start fulfilling its mission. X is not a junk drawer that is not suitable for special projects in other parts of the company. It is focused on trying to find ways to produce revolutionary products that are not only on the screen of the device, but are able to interact with the real world. By pushing Loon and Wing around the world, X will soon find out whether it can effectively hatch new Google - and whether it can put Alphabet in the forefront of the industry that has not yet existed.

But Alphabet's attempt to spawn a new generation of "moon project" companies raises two questions. Can this behemoth increase exponentially? Do we want it to grow exponentially?

Loon projectDevelopmentbehind

The Loon laboratory at X's mountain view headquarters is filled with several generations of fallen and leaky balloons. Loon is based on a simple idea - replacing ground base stations with high speed hot air balloons - which obscures a series of fascinating technical problems. In 2013, after a year of operation, these balloons still had a bad habit: they would burst or fall to the ground in a few days. (they carry parachutes to mitigate the impact of the load of electronic devices. The team will remind the air traffic control department of their landing.) Before a device like a "big bird" was realized - the launching process was just as simple as a group of children trying to fly a kite - a gust of wind could destroy the whole balloon.

Now, a custom designed box with all kinds of equipment can ensure the safety of the system during transportation. The key components are placed in a silver box, made of metallized polystyrene foam, which reflects sunlight and keeps it warm. A 80 foot plate scanner can check the tiny defects of the polyethylene, which can reduce the survival time of the balloon at 60 thousand feet from a few months to a few days. The mapping software can track balloons across the sea, and use machine learning to identify suitable airflow so that they can reach any place they need to reach. With these tools, the Loon team learned that the company could launch a balloon every half an hour and let the balloon stay in the air for 6 months or more.

At Loon's balloon forensics lab, Pam Desrochers used an 80-foot flat scanner to test polyethylene for minor defects and signs of wear after flight.

The forensics team used polarized lenses to detect defects that could reduce the survival time of balloons at 60 thousand feet from a few months to a few days.

This is the kind of development that X allows. For six years, Loon engineers, designers, and balloon recyclers don't have to worry about such problems as funds, income flows, recruitment of human resources, or who is responsible for the implementation of public relations strategies. They can use Google's machines to learn professional skills and X's "design kitchen" - a studio with an area of 20 thousand square feet to make prototypes of any mechanical device they can think of. They do not need to make detailed business plans, let alone generate income or make profits. They are allowed to fail again and again, and every failure can learn something.

Astro Teller, the X executive, described X as a place to make the world a better place, but he did not hide the benefits of Alphabet, including new revenue streams, strategic strengths and hiring value. Although he would not disclose the number of employees or operating budget of the project, he made it clear that no matter how much money you think X has spent, it is insignificant compared to the value it creates.

X director Astro Taylor defined the lunar landing project as the idea of solving major problems by proposing radical solutions and deploying breakthrough technologies.

The evaluation process of "lunar project"

Within X, teams pursuing various lunar projects are looking for their own path to failure, and they are similarly protected. Ideas are welcome as long as there are new solutions to difficult problems. They come from all walks of life, some from the brains of employees, others from Larry Page and Sergey Brin. X, co-founders of Taylor or Google, who delve into academic papers and attend conferences, Look for the seeds of projects that are expected to grow into giant trees. An unnamed project came from a researcher's NPR interview.

Regardless of where ideas come from, many of them end up with rapid assessment teams. This small team meets several times a week, not to promote ideas, but to refute them. "the first thing we have to ask is: can this idea be achieved through the technology that will emerge in the near future? Can it solve the right part of a real problem? " Phil Watson, who led the team, said. Breaking the laws of physics means nothing. "you'll be surprised at how many ideas about perpetual motion have been put forward." He pointed out.

These conferences have the atmosphere of unrestrained thinking in the smoky dormitory room, and the harsh and harsh atmosphere of thesis defense. The team has considered a variety of ideas: generating energy from Avalanche (infeasible), installing a copper ring around the north pole to make use of the earth's magnetic field to generate electricity (too expensive) and to build maritime ports to simplify transport logistics (which leads to huge regulatory trouble). They have debated about the development of stealth devices. It seems to be technically feasible. "We've been saying that we should do it, because it's cool - no, we can't do it, because it will bring more trouble, and it can't solve any real problems." Watson said, "it will bring great convenience to the lawless elements."

The idea of passing the first evaluation was taken to the Foundry stage. At this stage, the person who leads the fledgling project needs to solve the problems that run their project, something engineers have been reluctant to do. The leader of this phase is Obi Felten, who came to X.X. in 2012 after several years in charge of launching Google products in Europe.

In his first meeting with Taylor, ferton learned all the secret projects that X was brewing, including balloon networks and delivery drones. She began to ask questions that people would usually ask if they were responsible for product launches. Is it lawful for a balloon to enter the airspace of another country? Are there any privacy issues? Will you cooperate with telecom operators or compete with them? "Taylor looked at me and said," Oh, no one is really thinking about these problems. They are all engineers and scientists. We are just thinking about how to make balloons fly.

Any idea that can pass the scrutiny of a quick evaluation team will enter the Foundry phase led by Obi-Felten. At this stage, the person who leads the fledgling project needs to solve the problems that run their project, something engineers have been reluctant to do.

Foundry uses this strong interrogation technique to discover problems that may ruin a project before X can invest a lot of money and time. Take the Foghorn project of making a carbon neutral fuel with seawater as an example. This technology is amazing, and the problems to be addressed are significant, but two years later, the team realized that they had no way to compete with gasoline on cost - and to rely on technology that was more in the research phase than in the development phase. X killed Foghorn and gave everyone in the team a bonus for them to find new projects. Foundry is meant to ensure that inappropriate projects are eliminated as quickly as possible.

This is based on a simple premise: the sooner you kill an idea, the quicker you can put your time and money into the next idea. Trying to change the world and create huge new companies means avoiding traditional signs of progress. Exposing what you are doing and likely to fail is the only way to succeed. Because once it is completely killed, you can go back to do the next thing - something that might succeed.

Three major requirements of the project

Any project that wishes to meet the requirements of X must meet three requirements at the same time. It must be a major problem. It has to put forward a radical solution. It has to deploy breakthrough technology.

The definition did not appear until 2010 when X began to take shape, using it to separate delivery drones from cloaks. The effort began with an experiment in which Larry Page asked Sebastian Thrun, a Stanford computer science professor, to build a self-driving car for him. At the time, Trent knew the technology as well as anyone: he led Stanford in the 2005 DARPA Challenge. The event was a 132-mile full-automatic driving race in the Mojave Desert outside Prim, Nevada. DARPA held another more difficult race, the City Challenge, in 2007: getting cars moving In a simulated city, There they must obey traffic regulations, cross intersections, and stop cars. Stanford ended up in second place (Carnegie Mellon took the top spot) when Trent, who was already working with Google, came to Google to become a full-time employee to help develop Street View services.

The challenge of DARPA has proved that the car can drive automatically, but the agency no longer played such a game. At that time, the US car manufacturers were concerned about how to survive in the economic collapse rather than developing technologies that could destroy their business. Google is a software company, but it has a lot of cash reserves, and the idea of driving unmanned vehicles to the market is obviously hopeful to save life, generate new revenue streams, and extend Google's tentacles to other fields.

So Trevor quietly hired a team to focus on young people with their own minds, rather than leading scholars in the field. Many of those young people took part in the DARPA Challenge. (among them was Anthony Lewandowski, who was later accused by Google of stealing trade secrets, leading to a bitter legal dispute between the two companies that settled in February.) Page has created a challenge for the team: to select 1000 miles of California roads where he wants to develop vehicles that can drive themselves. Trent's team called it Larry 1000, and they finally completed the task within 18 months.

The creation of X

For Google, entering the real world is a completely new attempt. Its interest in projects beyond the core search business has spawned Gmail, Google maps, and Google books - all cool products, but all still belong to the software. The autopilot scene of TOYOTA Prius on the streets of mountain view inspired many possibilities, including more non software projects.

But the autopilot fell on the head of Google. The search for other equally difficult, complex and valuable questions requires some kind of infrastructure. Paige Jean Thrun became the first "other director" of the company to take charge of everything that does not meet the expectations of investors for Google. Because he wants to focus on the driverless car team. After 2012, he focused on his online business.educationThe start-up company Udacity, so his co head Astro Taylor came to steer a ship whose purpose and direction were unclear.

In his early communication with page, Taylor wanted to solve the problem of direction. "I said, 'are we an incubator?'" Taylor said. Not exactly. They are not the center of research. They are creating new businesses, but this does not convey the right meaning.

Finally, Taylor found an unexpected word. "Are we going to be on the moon?" He asked page. "Yes, that's what you're going to do." Page answered.

Creating a research department to create pioneering products is a mainstream practice for companies whose values are tied to their ability to innovate. This tradition dates back at least to 1925 by AT.

This breadth is key to the success of Bell Labs. You can't predict what the next breakthrough technology will look like, so there's no need to ask for a detailed action plan. "its leaders can accept that the goals of the project are not clear." Jon Gertner wrote in Creative Factory: the Great Times of Bell Labs and American Innovation, "Bell Labs staff will explore anything about human communication. Whether it's based on wire or radio, or on recording or visual images. "

However, Bell Labs operates within certain boundaries. Its most valuable tool is basic research: its scientists spent years exploring the basics of chemistry, physics, metallurgy, magnetism, and more time searching for promising discoveries. Although "human communication" is a wide range of areas, their work does not extend far beyond the ascent of AT

The Palo Artaud Research Center (PARC) of the Xerox Co (Xerox) is the first great innovation laboratory built by Silicon Valley, and its researchers stand out not because of their scientific breakthroughs, but because they have the ability to apply their original technology to new items that have never been considered. Mark. PARC created laser printers and Ethernet in the early and early 80s, and led the transition from a punch to a distributed interactive machine, the personal computer, and laid the foundation for modern computing.

But the best known in Silicon Valley, Parc, is that Xerox has failed to build on that work. The lab pioneered a graphical user interface-using the mouse to control icons on the screen-but Jobs eventually brought the interface to the public. Henry Chesbrough, who studies corporate innovation at the Haas School of Business at the University of California, Berkeley, points out that Xerox bosses are not wasting great technology; they just don't see how it relates to them. "Xerox is looking for something that fits the business model of photocopiers and printers."

X gives its employees almost unlimited authority and huge financial support to create products and services that the previous laboratory may have never found - or discarded products and services. Instead of doing basic research, it relies on other institutions, mainly government and academic institutions, to create tools that can be used for imagination. It does not depend on having the smartest people in the world, and is willing to find promising ideas and try to introduce them. Most importantly, its role is to expand the scope of Alphabet's business rather than improving existing businesses. The Baer laboratory, which won many Nobel prizes, is valuable to its owners because it makes telephone calls better and cheaper. Xerox's shareholders are grateful to the Palo Artaud research center because the laser printers it invented earned them billions of dollars.

X did not make these mistakes, because its duty is not to make search better. Its duty is to ensure that the Mothership of Alphabet will never stop expanding.

X's largest project

From this point of view, X's biggest project is not to develop autopilot, to create a balloon network, nor to conceive of unmanned drones - but to build a department designed to build these innovative companies. Its love of failure and the love of everyone's thoughts is to achieve this goal: if you are not constantly failing, or even foolishly failing, you are not working hard enough.

This is beneficial to Alphabet, beneficial to those who like the concept of unmanned vehicles (especially those who can't drive), beneficial to those who like to follow the idea of keeping track of health with noninvasive wearable devices, beneficial to those who enjoy the idea of immersing themselves in the Internet world in the dark corners of the world. People who like the idea of hamburgers and toothpaste that do not affect traffic and cause more carbon dioxide emissions.

But, through Google, Alphabet has actually had a huge impact on our daily life: how we talk to others, where we get the news, when we leave home to avoid traffic jams. For most people, it is good to get mailbox service, map service and search engine that can search almost anything. X tries to expand this influence by jumping out of the virtual world. Critics have been saying that Google is a monopoly. Now imagine that its rule extends to our cars, the food we eat, the goods we order, the health of our bodies - how we connect to the Internet. Google has already had a huge impact on our life without mobile phones. Are we ready to let it into other parts of our lives?

Behind the development of Wing project

Andre Prager pushed a cart full of junk into the room. The cardboard, which is mainly cut into small pieces, is mixed with bags of plastic debris and plastic endings. Pratt has previously worked on Porsche's car engines. In his spare time, he built a jet chainsaw and a turbocharged Vespa scooter. He is now a mechanical engineer for X's delivery drone program Wing. It was loaded with his various failed carts.

Wing Mechanical engineer Andre Plaget (left) is responsible for building a distribution system to simplify the process of getting packages from drones as much as possible. He and Trevor Shannon cut down on moving parts as much as possible. "our measure of success is how disapprovingly people see the device." Said Prager.

Shortly after the launch of the Wing project in 2012, the team realized that no chance of vertical landing on the ground would waste a lot of energy. Instead, they decided to let the UAV hovering in the air and somehow drop the parcel to the ground.

The first attempt of the team is a wire - based system in which the package is connected to a rope that can solve the cable on the UAV. "That sounds like a good idea because it's too simple." Platje said. But it quickly proved to be a complex mess: it was a great pain to roll things up. Each package needs its own system, because the cable is wrapped around the package, which makes it difficult for the customer to have a decent experience.

They tried to build a less complex mechanical system in the form of pencils and cabinet doors. (to show me, Prager took a prototype from the cart: a square piece of cardboard with a broken pen, a drawing pin, and a straw strapped to it.) Not at all-the package is always unhooked, the hook unhooked and reattached, or something breaks. "and then we said, what if we could do that without having to use any moving parts?" Another mechanical engineer in Australia, Trevor Shannon,videoThe meeting was expressed. Wing was tested in Australia.

When Wing teams test prototypes, they use the simplest tools to test new ideas: cardboard, tack and tape.

Thanks to a design like this,WingIt is ready to become an independent company and to distribute UAVs in the real world.

This idea led to their current design: the size and shape are similar to that of a small potato, with a dent on it. It is easy to fix with hands. When the load falls to the ground, the weight of the hook will naturally pull it down. There is a "anti jaw" to prevent the hook from re joining the package. He didn't mind its flatly appearance. "The standard we measure success is how disregarded when people see the device." He said.

The goal of Wing is to make it easier for people to get the goods they want, so that they do not have to spend time driving to buy. Since 2014, Wing has been conducting pilot projects in various parts of Australia, first in Queensland and then in the capital Canberra. The company began offering UAV delivery services to ranchers in remote areas, and is now preparing to start flying in the suburbs near the city. At present, the small packages it shipped from customers can be ordered from the large chain drugstore Chemist Warehouse and the fast food restaurant Guzman Y Gomez.

However, the real obstacle to the distribution is not the distribution system, nor the technology: in recent years, the battery and air control systems have made enough progress to allow a batch of delivery unmanned aerial vehicles to take off. It is how to deliver safely, especially in the United States and Europe, where airspace is crowded and strictly controlled.

So in 2015, the team began building a driverless air traffic management system, connecting all its drones and giving each drone a specific passage from the start to its destination. "We are trying to build freight trucks and roads for them." Adam Woodworth, who will be chief technology officer after Wing's independence from X, said. The difficulty with the project is that it's not just developing a system to track drones, but that everyone in the air is running the same system. Wing is working with the Federal Aviation Administration (FAA). And some parts of its system have been open-source so that other companies can develop interworking systems.

As Wing leaves X to become an independent company, its leaders-Chief Technology Officer Adam Woodworth and CEO James Burgess-face a real world where failure usually only means failure.

The interesting part of this problem is that it is not the kind of thing X aims to solve. It does not require engineering design, prototype design, or crazy brainstorming. It needs to be careful with regulators and competitors to communicate closely - for these entities, success means doing things well at one time. If Wing fails to do this, its long-term survival will be questionable.

The meaning of failure will be different

This marks a change that the new company will have to accept (Loon): graduation from X means a different relationship with failure. These companies are becoming successful companies in the traditional sense, to provide real services and provide real value to attract real customers.

The Loon balloon has flown nearly 20 million miles. Loon provided networks to Puerto Rico after Hurricane Maria and Peru's massive floods. Now, it has to do something more difficult. "it's time to get out of the nest." Alastair Westgarth, a senior telecoms industry figure, said. He arrived at X a year and a half ago and will be chief executive of Loon. His job will be to enter into partnership agreements with telecom companies around the world to integrate his balloons into their networks and provide networking services to their customers. Westgas points out that it is essential to remain bold and constantly promote innovation. "but at the same time, you don't want to risk survival."

Loon's chief executive will be Alastair Westgas, a veteran of the telecommunications industry who will be in charge of finalising cooperation agreements with telecom companies around the world to deploy their company's balloons into their networks. Provide network services to their customers.

In the real world, failure is failure. Slowly, your desire to die will become a survival instinct.

X will pay close attention to Loon and Wing, but its attention will soon shift to finding new "moon landing projects" to replace them. Whether Internet balloons and delivery Drones will dominate the sky or crash on the earth will take years to test. It will take a long time for us to find out the answers to X's failure, Tinker, improvement and introduction to others.

But back at Winnemucca, the balloon was rising steadily. It entered the desert, spent a night in the area, then flew to Denver, then Nebraska. Nick Corley told me that three balloons fired from a Puerto Rican base a few days ago were in the area. My eyes kept going back and forth until, guided by Cory, I found the smallest and whitest white dots moving quietly 62500 feet above my head. That's about 0.005 to the moon, and when you think about it, it's not that far away.

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments