NotCo - Copying Cows
Kaden Gulamani | Aug. 22, 2022
I love from time to time a good animal product in my diet and most of the world agrees with me. 350 million tons of meat (not including other animal products) are consumed every year, a number that has more than doubled since 1988 and is only going to grow faster.
Yet this trend of a meat-rich diet is only found in developed countries and is not only high in cost but highly inefficient. For reference, if everyone shared the meat-rich diet of Americans, there would only be enough space to cultivate livestock to feed 2.5 billion people, almost a third of the current world population. This trend only becomes more clear as the facts are divulged:
- It takes on average 54 calories of fuel to produce 1 calorie of protein, whereas it takes 2-3 calories of fuel to produce 1 calorie of protein in soybeans
- Over half of the world's food harvests goes to feeding livestock to produce animal products. HALF!!!!
- One hectare of beef or lamb feeds 1-2 people on average, while the same area feeds 19-22 people when repurposed with rice or potato
The list goes on and on, with livestock farming contributing to climate change and even water scarcity. Yet we can't suddenly force the whole world to give up its thirst for animal products, it's just not feasible. So how do can we effectively reduce the consumption of these products?
The main consensus among the world has been to create plant-based alternatives. Plant-based food has also become recently mainstream, with the mass marketing and adoption of Beyond Meat or Impossible Meat. This new trend is proven to combat climate change and does so through various means. Whether it's lowering processing energy costs, reducing the need for factory farms, or simply reducing methane emissions.
Yet plant-based alternatives require large amounts of research funding to fully replicate all aspects of meat. This includes the fattiness, the color, the texture, and various other criteria. Yes, companies do eventually figure out a broad formula to apply for some of these goals, however, one Chilean start-up believes it can apply machine learning to increase the efficiency of this process.
NotCo, founded in 2015, has since been acquired and partially owned by Bezos Expeditions, Jeff's own little for-fun investment firm. They originally used their technology to make more sustainable milk, aptly named Not Milk. Currently, Not Milk is widely available in the U.S and Latin America, but they also have produced Not Mayo, Not Ice Cream, and are breaking into the Not Meat market, replicating beef and tuna. One of its most successful products, Not Mayo, outsells most mayonnaises in Chile, and it's the 3rd most popular mayo brand in the country.
So how exactly did NotCo get to this point?
Its founders, Matias Muchnick, Pablo Zamora, and Karim Pichar, invented a state-of-the-art chef named Giuseppe. Giuseppe is an AI chef that has since become very sophisticated, trained to be able to emulate a particular taste, color, or texture of a given food. As outlined in UPenn's Digital Innovation blog, "data to train on can be made available by chefs attempting to make plant-based substitutes for foods or mixing random foods together and describing the qualities". However, the core value that Giuseppe brings to our world is its machine learning capabilities. As more recipes are tested and reported, Giuseppe learns and is able to maintain an extensive database to further predict chemical interactions when making a new recipe.
With each attempt, Giuseppe becomes better at guessing 100% plant-based recipes. This extraordinary process and its potential are evident when comparing the 10 months Giuseppe took to make Not Milk 2% versus the 2 months it took to create Not Milk Whole. Although Giuseppe isn't able to fully comprehend the complexity of the human palate at the moment, its recipes are a great baseline that professional chefs can build upon with preparation techniques. This can be seen starting at 12:40 in this youtube video, part of Youtube's "Age of A.I" series.
Giuseppe creates its recipes by identifying key chemical features in foods and applying its ever-expanding knowledge of plant products to figure out key ways to replicate those gastronomic formulas that our body recognizes in say a bite of burger or in a spoon of mayonnaise. The Age of AI series highlights this, by showcasing Giuseppe's findings that a chemical in cabbage emulates lactose, an analog of the protein commonly found in dairy products. Who could have expected red cabbage to be the main ingredient in replicating mayonnaise?
NotCo holds a promising future, and as Giuseppe's sensitivity to the human sense evolves with its database gathering more and more information, NotCo will possess a powerful tool to custom design plant-based food.
As my first blog post, NotCo was a pleasure to conduct research on and is such a fascinating company with a true one-of-a-kind technology. This is the beginning of a journey where I hope to expand my frontiers and learn about the ever-expanding world of biotechnology and its many applications.
I will not be doing an extensive bibliography as I quote everything directly taken and as this is a personal exploration, I hope the links to where I gathered my information are sufficient!
Feedback is appreciated! Expect a new start-up every Sunday (Start-up Sunday is my preliminary name, 100% original)