Of course, San Francisco is no stranger to robots at restaurants. I’m personally familiar with both coffee robot Cafe X and the ROBOJuice smoothie-making robot at the San Francisco Metreon. But neither offer a full meal — and full-service restaurants that do employ robots tend to still need humans to fulfill some tasks, such as taking orders. 

Mezli also isn’t quite fully autonomous. Real humans do the prep work for the Mediterranean menu (created by chef Eric Minnich of Michelin-starred Madera in Menlo Park) in a central kitchen offsite. 

It’s also more of a vending machine than it is a full-service restaurant — robots aren’t waiting tables. But from the moment you place your order to the moment a piping hot bowl is in your hands, the robots are in charge. 

Eoin Treacy's view

With a law proposed in California to raise only fast-food salaries to $22 an hour, the message is clear; innovate or die. Wage growth incentivises innovation in labour saving devices. That becomes more urgent as inflation pushes up the cost of everything else too.

Travis Kalanick (Uber’s first CEO) is building a network of dark kitchens around the world and focusing on delivery. That’s also part of Uber’s business model. The share is testing the sequence of lower rally highs and the region of the 200-day MA at present. It will need to sustain a move above $35 to confirm a break of the downtrend.

Deliveroo is back testing the upper side of its emerging base formation near 100p. It will need to sustain a move above that level to confirm a return to demand dominance.  

Delivery robots are being actively tested and one rolled through a crime scene in Los Angeles this week. They are not yet fully autonomous, but a lot of venture money is going into betting they will be some day.

This snippet from Tomorrow’s World in the 1970s highlights how much visibility of the future was possible at the time, and how long it took for economies of scale to catch up. Even back then, they were still making predictions about what would be possible within a five-year window.

The reality is commercialization often takes decades and that is when the technology actually exists. With autonomous vehicles, the technology does not exist. It has to be created from scratch and that is a major technological challenge. A massive leap in how computers gather, understand and act on real time information will be required for full autonomy. That’s a very tall order.