This week, I attended the one-day Food Robotics Summit in San Francisco. The event featured speakers from many of the companies working on automation for cooking, restaurants, and delivery, and some of the attendees came from leading retailers.
The Spoon, a Web news portal covering the food tech revolution, hosted this event. It also runs the annual Smart Kitchen Summit in Seattle. As an overview, it prepared this market map of the food robotics landscape:
Among the attendees at the summit (also called ArticulATE) were techies and executives from a variety of startups, as well as from Walmart, Albertsons, Google, and Sony.
The panels were moderated by writers from The Spoon. They discussed grocery delivery and fulfillment methods, robotic delivery considerations for the last mile, and 24-hour storefronts and point-of-sale automation.
The panels also covered safe robots for food preparation and artificial intelligence applications. In addition, the speakers shared their experiences and some really interesting perspectives about the challenges the industry faces.
One of those perspectives came from the inventor of the original Sony Aibo robot dog. Masahiro Fujita (left) now heads Sony’s AI Collaboration Office, which is focused on futuristic consumer products.
He talked about Sony being in the consumer products business for the long haul and showed a video of a concept kitchen using all sorts of new technologies:
Another perspective came from Vincent Vanhoucke, principal scientist at Google X Robotics. He talked about AI, machine learning, and deep learning as it relates to vision systems and grasping. Vanhoucke emphasized the messiness of most human-robot interactions across all verticals and said there is a lack of any really good grasping/perception choices.
Trung Nguyen, vice president of e-commerce at Albertsons, talked about his company’s plans to provide fulfillment and pickup (or delivery) of online grocery orders. But he brought up an issue that all of the VCs and big chain representatives mentioned: the ability to scale and to provide support. Albertsons plans to roll out its fulfillment and pickup solution to 3,000 stores.
Walmart just did the same thing with orders to Brain Corp. and other vendors. I met and talked with the guy from Walmart who placed those orders, and part of his due diligence was to see how the supplier planned to meet those massive quantities and provide support. Panelists in the venture capital panel said they’re always looking to ensure that prospective startups can scale and support those new numbers.
LG Electronics also announced that it is working with South Korean franchise owner CJ Foodville to develop robots for its restaurants.
Data and design for food robotics
There were panels on data collection and use, machine and deep learning, novelty versus true problem-solving, and how slow that process is. One example of that slowness was described by Alex Vardakostas, CEO of Creator (previously Momentum Machines). The company is creating fresh hamburgers to order at its restaurant in San Francisco.
He described the eight years of design changes that were required for safety purposes. Vardakostas also told about how Creator’s product changed from a backroom production engine for fast-food producers to a chic in-store colorful marvel of color and function and, after four and a half minutes, a tasty hamburger. Martin Buehler, who previously worked at Disney and Vecna, is the head of engineering at Creator.
In line at the Creator restaurant were at least 10 of the people from the previous day’s food robotics summit, including Masahiro Fujita. The hamburger was good and not expensive. (Note: The restaurant is only open during lunchtime Wednesday through Saturday.)
There were demos and/or presentations from CafeX, which had a booth right across the street; ChowBotics, which had a salad-making kiosk at the summit; and Wilkerson Bakery, maker of the BreadBot. Delivery robots at the show included those from Kiwi, Starship Technologies, and Bear Robotics.
In addition, there was Miso Robotics, whose Flippy robot makes hamburgers. (For LA Dodger fans, next time you go to Dodger Stadium, look behind the chicken counter at FD 7 — through the glass partition — and you will see a couple of blue shrouded Miso robots grilling chickens.
More discussions at the Food Robotics Summit examined interesting uses of market data, such as in restocking kiosks and informing management of productivity metrics, or in machine learning from robotic operations.
Although this conference focused on the restaurant, food production, and presentation side of the industry, it was fascinating and fast-paced. The event was worth attending, even though there was no mention of robotics on the agricultural side of the industry. Nevertheless, the challenges presented — particularly grasping, perception, scaling, safety, and support — seem universal for robotics developers and vendors to address.