They call it "the hive" or "the grid". Or sometimes simply: "the machine". It's a huge structure that fills a warehouse on the outskirts of Andover, a small, quiet town in the southeast of England. It is impossible to cover it at a glance, but standing on a maintenance walkway near the building's beams, you look over what appears to be a large chessboard, populated entirely by robots. There are more than a thousand of them, each one the size and shape of a washing machine, and they circulate, day and night, moving food. Their job is to be cheaper and more efficient than humans, and they are very good at that.
The hive machine is the brainchild of Ocado, a British online supermarket that has made a name in recent years by designing highly automated warehouses and selling the technology to other supermarket chains. When fully operational, Ondo's Andover operation will be the most advanced so far, processing 3.5 million items or around 65,000 orders each week. It is also a perfect example of the wave of automation that slowly hits countries around the world. The tasks carried out by Ocado bots are so basic that they are best described with simple verbs: "lift", "move", "order", and that means that they exist in various forms in a wide range of industries. And when the price is right, someone will want a machine to do those jobs too.
Although robots are the main attraction in Andover, there are still many humans stumbling. One of them is Ocado's technology director, Paul Clarke, who joined the company more than a decade ago and has been entrusted with the development of its automated operations.
Explain that the objective of Ocado is to "disrupt itself" to continually update its technology so that it can not be overtaken by the competition. In the oldest and now outdated warehouses of the company (which, according to Clarke, are still some of the most advanced in the world), groceries are processed more or less linearly. Deliveries are unpacked in boxes; the boxes are placed on conveyors; and the transporters take the boxes to the shelves, where the human "gatherers" take what they need to complete the orders of the clients. The new paradigm, however, is about using space in the most efficient way possible. The items are still placed in boxes, but those boxes are now stored in large quantities, up to 17 boxes high. Its position in this pile seems to be random, a box of shavers next to the fillets of cod, for example, but it is decided algorithmically; with frequent access items placed at the top and rarer purchases near the bottom. In addition to this treasure, robots do their job.
Each of the bots has a central cavity and a set of claws used to grab boxes and place them inside, as an extraterrestrial abduction in a supermarket aisle. You can then move the box to a new location or place it in a vertical conduit to a collection station. In these stations, human employees take the items they need from the box (a screen in front of them tells them what to wear) and place them in a shopping bag in another box . Both boxes are sent back to the network, to be refilled with shopping items or moved to the delivery bay.
Imagine a huge machine, with groceries at one end and purchase orders at the other. The humans carry out the unpacking and packing, while in the middle, the robots sort and reorganize this vast inventory 24 hours a day.
Individually, bots are not smart; they do not make decisions for themselves. But all its actions are coordinated by a central computer. Clarke explains that this system means that robots can be used in the most efficient way possible. For example, when working as a team to quickly search the stack and recover unusual elements. "If you want to choose a typical Ocado order of 50 items, they will help each other," he says. A group of robots can meet in groups, divide, "and choose that order in a matter of minutes." In a traditional warehouse where items are scattered on distant shelves, this process can take hours.
Besides the impulse to speed, the grid machine has the advantages of being scalable and modular. If customers want to increase the size of their operation, they only add more boxes and robots. And if any individual robot breaks down, it does not matter, because any of the other bots can do their job; all are interchangeable This means that Ocado has only one robot to "design, evolve, manufacture and support", Clarke adds. "And that leads to economies of scale, because we have reduced all that mechanistic diversity to a common component."
This sales argument obviously has something to do, since in the last year, Ocado has made deals with supermarket chains in France, Canada and Sweden to update their warehouses. Such agreements should make it easier for these companies to offer online purchases (the UK is a relatively early adopter of this trend) and will help to avoid fears that technologically experienced rivals will enter their territory. See, for example, Amazon buy Whole Foods.
But while the focus here is on technological advances, in more and more automation, we must not forget that in the middle of these machines – metaphorically and literally, in the case of the Andover warehouse – there are humans.
You may have seen some pretty impressive headlines about artificial intelligence defeating humans in this or that task, but it's worth remembering that nothing catches a robot like a bag of oranges. They just can not deal with that. The bag moves in many strange ways, there are no obvious parts to grab, and if you squeeze too much you end up with orange juice. That's why Ocado still employs many humans.
They work in several key positions in the warehouse, which are also, if you know what to look for, technological bottlenecks. Robots still can not decompress the wide variety of bulk deliveries that arrive at Andover every day; nor can they move pallets quickly around a warehouse occupied by forklifts. And although they still can not handle bags of oranges (or other delicate or irregularly shaped items), Ocado is working on a solution.
Placed, somewhat insensibly could be said, next to the picking stations manned by humans, there is an experimental position where a robot arm is learning to do what naturally comes to his fleshy colleagues. Namely, take items out of boxes and put them in shopping bags. The arm is equipped with a suction cup, which is ideal for holding objects with rigid and flat surfaces, such as cans and cardboard boxes, but still can not deal with more delicate items. For that, Ocado is developing a soft robotic hand that uses rubber fingers filled with pressurized air. Seeing him grab a lime is a disturbing experience, with his synthetic fingers curled around the fruit like pythons.
Neither the suction cup nor the rubber hand are ready for prime time, but Ocado says that robots like this one should be integrated into their stores in the next few years. And it's not the only company that works on the problem. Amazon organizes an annual "collection challenge" in which teams compete to create the fastest robot pickers. (They are competing for cash prize and prestige, but some also expect Amazon to choose them as an acquisition). Well-funded startups are also building their own solutions. An Embodied Intelligence call uses AI to create robots that learn by observing humans. Another, Kindred, uses traditional robot arms, but has human engineers who can operate them remotely using virtual reality when they get stuck.
It may seem like a great effort to solve a relatively trivial task, but the better the robots are to mimic the ability of humans to manipulate delicate objects, the greater their use in almost any other manufacturing process you can think of. A recent study found that the field of logistics has been one of the first to adopt robots and artificial intelligence, mainly because the tasks involved are relatively routine and, therefore, easy to automate. But experts say that the technology that is incubated here will be adopted by other sectors.
"[Picking] is a problem that people try to solve in many different use cases," explains Euan Cameron, an analyst at PWC The Verge . "And these solutions will be collected and transferred to other industries."
Estimates of how many jobs could be lost for robots and AI vary, but a recent OECD study suggested that about 14 percent of occupations in developed countries (such as the US, Canada, and Japan) ) are at high risk. And a large part of these are found in logistics and related fields, such as storage, distribution and compliance. A PWC report found that in the UK alone, more than 10 million logistics workers are at risk of having their jobs automated in the next 15 years.
When I ask Clarke if the ultimate goal for the company's factories is to have no human workers at all, he offers a balanced answer: "In theory, but it's not something we can reach in the foreseeable future" . He points out that even in industries that have invested heavily in automation, such as automobile manufacturing, there are still many humans involved. "And for us, it's exactly the same road we've been on since the first day: look for the next thing to automate, either by placing plastic bags in boxes or moving goods in our warehouses." We start with the obvious and move forward to automate what next and the following, you never get to the end. "
Certainly there are jobs in the Andover warehouse that will be resistant to automation for a while. Repair robots that fix robots, for example. On the maintenance walkway, I see a huge rescue car designed to be taken to the network to grab smaller robots when they break down. When I ask engineer Dean Tharme how to recover the rescue car if it breaks, he simply responds: "with difficulty".
Tharme, 29, has been working at Andover since January. He used to be an electrician before being hired by Ocado to service the robotic herd of the warehouse and is one of several engineers whose work benches line the perimeter of the network, huddled next to scattered kitchen facilities and a control station. (To watch the robots, not the humans). Here, Tharme and his co-workers replace the broken wheels, re-weld antennas and, in general, make sure that the robots can continue to transport in trucks. Broken bots are archived at one end of the gangway and re-open to the grid in the other. It reminds me of having seen the sheep being sent back to the pasture after shearing.
Tharme says he takes a lot of pride in his work, and prefers it to his old job as an electrician. "Every repair he does has his name, so he wants to make sure that the corrections he makes are good," he says. "Sometimes he repairs one, sends it and comes back in just thirty seconds because he did not do it right, it's a disappointment." Compare the bots with a "children's fleet", but he says he does not have much personal connection with them. "Even though you start to recognize some of them," he says. "Specific numbers that have had failures in the past, you know they are a problem."
Being a robot repairman is a race with a future. While estimates of whether automation will be a net job destroyer vary, economists agree that technology is likely to polarize the labor market, dividing labor into two fields: high-paying, highly-skilled jobs at one end, and little remunerated, low-skilled jobs on the other. Think of the contrast in a company like Uber, which employs computer engineers in Silicon Valley with six-figure salaries, but also hundreds of thousands of drivers who work hours for uncertain rewards.
It is difficult not to see a similar dynamic in the warehouse of Ocado, although with a less extreme separation. The company says it never destroyed any work; In fact, he says, it is the opposite. Because automation has been part of your business since the beginning, it has created more than 14,000 jobs that would not exist without robots. But that does not mean that your future stores will have so many people employed, or that there will be no division between your employees who pack purchases and those who design the robots.
And in the same way that the human employees of Ocado are contributing to the operation of a machine much larger than them, the company is only a part in the vast engine of technological and social change that we refer to as "automation" .
When I talk to one of the warehouse workers sitting outside for lunch, he tells me he is not worried about a machine taking away his job. He says that the work is pretty boring, and in any case, he has seen the experimental robot arm in action and thinks it will be a few years before he can take over from a human being. And after that, I ask: what happens when the technology is good enough? "Well," he says, "I'm training to be an engineer anyway"
Photograph by James Vincent / The Verge