Why you should let a robot cook your next meal

Arriving at Creator, a new restaurant located on the ground floor of an office building on downtown San Francisco’s Folsom Street, feels like walking into a catalog. Sleek, wooden communal tables with high white stools line one end of the room, with a bookshelf full of hand-picked culinary books against the wall and modern light fixtures overhead. On the other end, however, two large machines each the size of a small car form a 90-degree angle around a center counter, giving the whole space a retro-futurist aesthetic. It’s what you would have imagined a restaurant eventually looking like if you watched a lot of The Jetsons.

Those machines, with large transparent glass casings and ingredients in cylindrical tubes, are Creator’s burger-making robots. Each 14-foot device contains around 350 sensors and 20 microcomputers to produce the best, freshest, locally sourced cheeseburger that $6 can get you in America’s most expensive city. After trying the first one out of the machine that day, I can confidently say the claim holds up; it’s the best-tasting burger for the money I’ve ever had. The machine is also capable of making multiple kinds of burgers, with vastly different flavor, ingredient, and condiment combinations. In a way, it feels like the future of fast casual food.

“We wanted to design a device that meets nature where it is, and not make food conform to a robot,” CEO Alex Vardakostas tells The Verge in an interview. “We didn’t want something that would make one kind of burger. We look at this like a platform for recipes, and we wanted as much culinary creativity as possible.” Right now, Creator is only taking reservations for 30-minute lunchtime slots on Wednesdays and Thursdays through the month of July, and plans to do the same in August while it irons out kinks and improves its software and workflow. But, eventually, Vardakostas says you’ll be able to order takeout using the company’s mobile app, while workers with iPhones will stroll around the restaurant to take eat-in orders, just like an Apple Store.

Creator, formerly known as Momentum Machines, is one of a rising new type of automated restaurant, mixing the best of the tech industry’s software, robotics, and artificial intelligence skills with top-tier culinary expertise. The goal is not to automate away humans entirely, but to automate the portion of the restaurant experience that can be done better, faster, and be more cost efficient with machines. Creator joins companies like San Francisco-based quinoa bowl chain Eatsa, pizza-delivery company Zume in Mountain View, CaliBurger parent company and Miso Robotics investor Cali Group, and a smattering of up-and-coming locations around the country like Boston’s Spyce and Seattle’s Junkichi.

On one end, it’s about convenience; well-programmed robots can make food faster than people, while touchscreens and mobile ordering cut down on long lines and clumsy interactions with cashiers. “It’s not that people are catching up with this type of model. It’s that this model is catching up with the way people and society are tending to interact with companies,” says Eatsa CEO Tim Young. The Eatsa experience involves no human beings whatsoever in the standard ordering process and instead lets customers order from an iPad. Food is prepared partially by automated machines and partially by humans in the kitchen before being placed into mechanized cubbies a customer opens with the tap of a finger on the glass.

“Customers are interacting through their devices,” Young adds. “Customers are demanding this type of experience. The services and goods that people are purchasing, they are, in all walks of life, usually doing it through a device.” In addition to convenience, Young says automation is helping drive the kinds of efficiency gains and cost reductions that let customers buy healthier food for less money. (A typical Eatsa quinoa bowl costs around $7.) Young says the model also helps workers stay in food service for longer than the industry-standard six- to nine-month stints.

For Creator, Vardakostas is starting with automating only the burger-making process. There are still humans to take your order, prep sides like fries and coleslaw, and monitor the machines to ensure they’re not dousing the wrong side of the bun with mustard. Vardakostas, a former engineer, has spent the better part of eight years perfecting his vision for an automated burger joint that could make better food for less money, all while employing roughly the same amount of human beings as your standard McDonald’s franchise. Right now, he’s eschewing touchscreen ordering because he feels it’s too impersonal.

Vardakostas tells me, over four of the different Creator burger varieties he’s had the machine prep for us, that his ultimate aim is to create a superior customer service and dining experience, one that uses robots to help humans focus on creating better recipes and tastier food while maintaining a cheaper bill for customers. “One of the great things about our operation is that because it’s so efficient, we spend much more on the cost of ingredients, which really comes through in the flavor,” he says. “As an example, the beef is whole chunks of chuck and brisket. It’s steak, it’s pasture-raised — no antibiotics, no hormones.” He imagines the Creator experience involving order-takers interviewing customers about their desired taste profiles and burger-making proclivities, to help tailor the meal accordingly.

Beyond being home to Silicon Valley, there are other reasons why the San Francisco Bay Area is a becoming a hot spot for experimental automation in the restaurant space. For one, real estate is among the most expensive in the country, and San Francisco in particular has strict laws around opening chain stores. California minimum wage is also now $15 an hour as of last week, while state labor laws mandate that servers take home the full wage in addition to tips, unlike in other states where servers typically make a reduced hourly wage.

That creates an interesting predicament for restaurant owners and workers alike, with managers hesitant to hire new servers due to soaring labor and operations costs and servers unable to live in the city they’d be working in because of sky-high rent. (Studies have also proven that higher housing costs will increase the cost of local services, including dining.) As The New York Times reported late last month, a new breed of counter-service restaurants like Greek fast casual spot Souvla are responding to these economic forces by serving higher-end food costing in the double digits per meal, yet doing so without any wait staff. At Souvla, based in San Francisco, diners fetch their own silverware and fill their own water cups. It’s Chipotle, but with better margins and the veneer of a sit-down restaurant with 4.5 stars on Yelp.

“I think in a city like San Francisco, there’s two things that make labor more expensive: low supply of low-skilled labor because they can’t afford to live there. The other is minimum wages that feed into the same problem. Both of those things push up labor costs,” says David Neumark, an economist and professor at the University of California, Irvine, who studies wages and labor markets.

The reality is that modern restaurants in expensive cities are being forced to compromise, either on aspects of customer service like wait staff to achieve high-quality food, or on the quality of the food to achieve the speed and efficiency some customers demand. But another option is automation. Robotics and software advances have made it possible to produce more food at a higher quality and at a price competitive with fast food restaurants.

For full automation, you need to design that system from the ground up, as Creator did. Other companies like Cali Group, which invested in a California Institute of Technology spinoff that became Miso Robotics, are taking an incremental approach. The company worked with Miso to develop Flippy, a burger-flipping robot that uses computer vision and other artificial intelligence features to operate a grill. The robot is working now out of a Caliburger restaurant in Pasadena, California, where it’s been performing nearly 100 percent of the tasks a typical short-order cook is responsible for. Eventually, the company wants to roll out Flippy to more Caliburger locations, while Miso works on reprogramming additional robots to perform other restaurant tasks.

“The argument for the whole industry to move to automation is that, macroeconomically, demand for restaurant food is going up, and people don’t necessarily want to go into the restaurant,” says Cali Group CEO John Miller. He notes how with on-demand delivery becoming more popular, people are going to cook less while the costs of cooking that food should go down. “With that, the big opportunity for automation is the whole industry can have higher output with the same number of workers. That’s the big story — productivity per worker. If you can make more food with the same number of people to meet the increasing demand from the public, that’s what needed.”

And watching Creator’s burger-maker in action, I’m in awe at how such a machine could operate on a larger scale years down the line. It is effectively a fully automated, end-to-end burger assembly line running completely on code. Large tubes of fresh burger buns funnel the bread down a ramp on the right and into a toaster, where the bun is warmed until it has a golden brown crust on the insides. From there, the bun is laid flat on a bed of copper-colored metal teeth that push it down the production line by flipping end over end. The bun stops at a condiment dispensing station, and then one for toppings like pickles, lettuce, and cheese, all of which are cut and shredded on the spot by the machine itself. Finally, the bun receives its cooked patty from an encased grill on the other end. When completed, a human server takes the burger out, inspects it for quality, cuts it down the middle, and preps it neatly on a plate.

The whole process takes no more than five minutes, and there I am sitting in front of a $6 burger that, had I ordered it anywhere else in the city, I would have sworn cost at least $10 to $12. Vardakostas says when both of Creator’s two burger-making machines get up and running at full capacity, they’ll be capable of putting out 120 burgers each per hour. It’s that kind of capacity that will let Creator operate in a city as costly as San Francisco, with the roughly 50 employees it has now (including engineers and management) and the 10 or so people it will take to operate the restaurant on any given day once it opens full-time.  

An open question is whether models like Creator are the future for all of food service, or whether only boutique restaurants and venture capital-funded outfits will be able to do automation in the near term in a humane, sustainable way. It’s unlikely McDonald’s, which has rolled out touchscreen ordering at locations around the world, will start employing engineers, trained chefs, and robot operators or promoting its existing workers, though it certainly could years down the line. Yet as restaurant jobs become more complicated as more robots enter the equation, it will become inevitably harder for less-skilled workers to obtain those better-paying positions, especially outside of big cities like San Francisco with large technical workforces.

“There’s nothing new about automation. We used to have 80 percent of the workforce on farms, and now it’s 2 percent,” Neumark says. “If a machine can now do something at lower cost than a person, there is now some higher-value activity for that person somewhere in that economy.” Though he does say that with automation comes dislocation, meaning it may take laid off workers some time to find new jobs and that there won’t be that higher-value position available for every single worker.

Right now, seasoned automation companies like Eatsa and Zume have already started discussions with big brands about licensing their platforms, so we may start to see versions of robotic restaurants popping up around the country using a variety of economic models. “I think Chipotle is already talking about doing these sorts of things… Starbucks talks about [digital experiences] openly,” Eatsa’s Young tells me. “It’s far from dystopian. It’s what their customers are demanding.”

“We’ve been working on this end-to-end platform for about three years, and the intellectual property creation for that was even earlier. Anything that sounds simple in this is actually really complex to do well,” says Chris Satchell, the chief technology officer of Zume. The company uses robotics to automate portions of the pizza-making process, like pressing dough and spreading sauce, and the fleet of trucks that cook those pizzas en route to customers. All the while, proprietary prediction algorithms dictate how much pizza to produce and for which neighborhoods on any given day depending on order histories and inbound demand. “What I think you’ll see is big companies and small companies coming to specialist platforms like Zume, especially companies that have proven they can do this with their own first-party platform.”

Zume, like Eatsa and Creator, is dedicated to doing automation in a thoughtful, pro-worker manner, and Satchell is convinced the only way to “feed the world without ruining it” and to do so with healthy options and fresh ingredients is to involve robots. If you don’t, cost and efficiency pressures will always drive down quality, he says. Zume refers to its workforce as “co-bots,” as in collaborative robots that work together with humans. The company now employs about 150 people, and it provides health insurance and what Satchell says is a good salary for food service in Silicon Valley. “It’s not about how to replace everybody. It’s about how to elevate the humans that you have,” he tells me.

Recently, the Organisation for Economic Co-operation and Development, or OECD, studied high-income countries to better evaluate which jobs were automatable, and its April report found that 14 percent met the qualification. It provided a stark contrast to the widely cited and influential 2013 paper by Oxford University academics Carl Frey and Michael Osborne that predicted around half of all jobs in the US would be automated in the coming decades. Still, 14 percent across countries like the US, UK, Canada, and Japan equates to around 66 million jobs lost, and food service is prime among the categories of highly automatable positions. In the US, the report suggests around 13 million jobs will be automated, but it does not give a concrete timeline for when that will happen or at what pace it will accelerate.

Yet while fears of automation in the food service business may be top of mind, but it’s not very likely to happen in widespread fashion any time soon. As Vardakostas tells me, it took him and his team years of working on their own before they could court funding to even embark on the quest of producing Creator’s burger-making machines. Even now, the restaurant will have to wait months, perhaps even until next year, to open to the public because Vardakostas is not sure it can sustain the necessary hours and food output. One of the biggest bottlenecks at the moment is that the machines must be shut down every so often to ensure the grill can be properly cleaned.

Vardakostas’ biggest fear now, he says, is a day where both of the machines break down in some way, and Creator can’t make a single burger at all, let alone a $6 one in five minutes. Zume has yet a announce partners that will be using its platforms in the near future, with negotiations still ongoing. Eatsa shuttered its New York City and Berkeley locations last year, with the company explaining that it expanded too quickly, but it did secure its first third-party partner in Chicago-based Asian fast casual chain Wow Bao last November.

The general techno-optimism around automation is that it will produce better-paying, more fulfilling jobs for humans as menial tasks get taken over by robots and software. The pessimistic outlook is that AI, unlike the move from farms to factories, will accelerate so fast and cover so many industries that the rate of job loss will outpace society’s ability to find new positions for the unemployed.

The reality is likely somewhere in the middle. And yet ground zero of this movement is not in the boardroom of an evil corporation bent on maximizing profits, but in small experimental dining experiences drafted by startups like Creator. It’s easy to forecast how this could all go horribly wrong: a dystopian picture of mass unemployment and a robotized burger joint on every block. But for the time being, tasting this $6 cheeseburger made by a machine and listening to Vardakostas explain his customer service vision is enough to convince me that there is a bright future for robots in restaurants, ones that work alongside human beings instead of as their replacements.