A machine cannot smell that perfect cooking moment
- Scott Daniels: We Ate Well and Cheaply
- April 13, 2023
- 568
Some time ago, I saw a program in which people were sitting around a table in South Korea enjoying late night drinks and a few bowls of amazing noodles. One of them, an economist, said one of the greatest challenges of the coming century was how to support those workers who were displaced by technology.
He said there was a growing population of people who, through no fault of their own, had no work to do. They didn’t lack education or motivation. They had marketable skills and wanted very much to work. There was simply nothing for them to do as so much of society is supported by automation and artificial intelligence. Some sort of minimum guaranteed income, the economist argued, would be unavoidable at some future tipping point.
Faced with such a scenario, a young person might look for a robot-proof career, something which requires a human to accomplish, and those are becoming scarce as artificial intelligence is deployed in more and more industries. A.I. can even pass a bar exam, and with a limitless ability to scour for case precedent in seconds, even courtroom lawyers may become obsolete.
At least cooking requires a person with a nose and eyes. A machine cannot smell the perfect moment when vegetables are perfectly roasted just before they become charred and bitter. Chefs are safe, right?
Not so much, actually. A California company, CloudChef, garnered some coverage recently when they showed off a startup that replaces chefs with artificial intelligence and some modified kitchen equipment.
The A.I. bot can watch a person prepare a dish once, making note of temperatures and textures, then replicate the dish exactly an endless number of times. The company envisions a time when well-known chefs around the world would be able to upload their recipes, which could then be prepared in bot kitchens anywhere.
With enough buildout, you could get a perfectly replicated Gordon Ramsay beef Wellington in your own city, whether or not the chef had a restaurant within 2,000 miles. Each time the recipe was prepared, our random Ramsay example would get paid a royalty, and an army of cooks would be sitting at home with no way to make a living at their craft. Back in their kitchens, barcodes replace eyeballs, knives are automatically sharpened, and the metal and chip cook never trips over the rubber mats or incorrectly stores the tuna.
Human chefs, naturally, express serious doubts about dishes turned out by machinery. Food quality varies daily, they say. No machine can detect the adjustments needed to cover for such daily variables. How can a machine taste when a sauce needs more salt?
One of the great truths of cooking is that every time you make a dish, it is necessarily a different process. The ingredients you used last time will vary slightly this time. Humidity, range top quirkiness, tools, pans, all these things can make a difference. Replicating a dish as it was cooked one time, potentially years ago, in exactly the same way cannot be successful.
Except that it is. The A.I. software is able to tell if the leaves on a Brussels sprout are too brown or if the bacon has gone off. It can adjust on the fly to get the desired identical result.
I think this scenario is alarming, and just as the South Korean economist pointed out, we will have to adjust our attitude about those who are unable to find work in their field. If we employ A.I. to do everything from driving taxicabs to injecting vaccines, we can’t in turn blame the people who are displaced when they sit at home with nothing to do.
And then you see a great cook with nothing but a slab of plywood and a long pin roll out sheets of pasta dough the size of a sofa and so thin you can read a newspaper through it, and you know the thing machinery cannot replicate: love. Cooking for others is an essentially human endeavor, and we are wise to keep smarty pants computers and their barcodes out of the kitchen.