Algorithmic Kitchen

algorithmickitchen By Luis Rodil-Fernández

It is very likely that in the last few years you have come across various uses of the word algorithm. A term that until not long ago was confined to the bottom shelves of Computer Science libraries. The appearance of this word in popular media sometimes comes hand in hand with spooky stories of predictive profiling and invasions of privacy. Inevitably myths are forming around algorithms and popular lore often bestows supernatural powers to these mathematical constructs. We read about algorithms that can predict if we will be lucky in love. Algorithms that can predict when we will die. Algorithms that can tell when we are pregnant before anybody else knows. Or algorithms that can best decide how to spread our inheritance when we die. What strange powers are these and how did computers get to figure all this out?

In popular lore algorithm has become synonymous with computer thing that does incredible stuff. Very little more is known about how these mysterious and mostly secret sauces do the things they do. Sometimes it is because the algorithm effectively codifies a trade secret that has made someone very rich and he, because it almost always is a he, is not yet willing to tell us how he did it. But most of the times it is because nobody really offers any explanation whatsoever.

There is a certain sensation of powerlessness that comes from being confronted with the reality of these new digital gods. They pour day and night over our online clicks, our medical records, our car’s engine, our Facebook likes and our smartphone’s location, they are omnipresent and inscrutable in their ways. We can hardly know what they do or how they do it. They have access to our unconscious by analyzing our habits, our ticks and the things we do when we think we are not being observed. All this so that they can tap into our innermost desires, so that our weaknesses can be unveiled and targeted.

Being overexposed to digital media can leave one exhausted, feeling helpless and dependent. My favourite and most effective way of dealing with this extenuation is the kitchen. The kitchen is the sacred space of the analog, a temple to capricious cooking times, vague measurements, good-enough approximations and improvisation. Like a jazz musician one has to atune to sounds, smells and appearances to match the beat when the time comes to add the garlic, remove the rice from the fire or know when the veal is done just right. A place where efficiency and optimization do not mean much and not always result in a tastier meal.

Algorithms, when explained to those not in the know, are often compared to cookery recipes. As a kind of computational recipe that precisely defines the steps to process certain kinds of inputs into certain kinds of outputs. The comparison is moot however, because while a slight variation in a recipe might result in a more tasty dish, it is a precondition of the algorithm that it has to be implemented with great mathematical correctness to lead to the expected results. The world of algorithms also happens to be far more bizarre and counter-intuitive than the kitchen. Cookery recipes fall short of explaining algorithms that can learn and refine themselves over time or algorithms that can simply evolve out of a set of very few simple ingredients with no human intervention. The analogy of the cookery recipe can only stretch so far to explain the digital world. Algorithms are not bound to the physical reality of foodstuffs, they are a different kind of thing, so the metaphor doesn’t always stand.

algorithmickitchen2I love writing software and I love to cook, the convergence of these two activities is, to me at least, a feast in all senses. Observing the gap between our poor understanding of what algorithms are actually doing with our data, the ineffective analogies used by popular media to explain them and the growing fears and myths surrounding these digital gods that live in black boxes in some refrigerated datacenter. I came up with the idea of designing a workshop where people can learn the basics of some algorithms that are heavily used in industry and apply this knowledge to a database of cookery that has been assembled by an algorithm that I myself designed. Perhaps by building a basic recommender algorithm using your favorite food ingredients you can get to better understand how Amazon knows that people that bought this were also interested in that. Perhaps building a simple regression model helps you to better understand how your insurance company is using your data to make forecasts and maybe use that knowledge to compute a menu that will better regulate your intake of sugars.

The kitchen itself is a technology, a well proven, constantly evolving and rich technology from which the keen eye can draw many productive analogies for a deeper understanding of the digital world that seems so far out of reach to non-geeks. The recombination of elements in the Algorithmic Kitchen’s database can result in surprising dishes, that don’t adhere to cookery tradition or intuition. However nobody can tell whether it is nasty or delicious until we feel it in our tastebuds, that’s why we close the workshop by cooking and eating together the results of our algorithms, a most primal form of social networking.

Algorithmic Kitchen is a project by Luis Rodil-Fernández in the cadre of Algo Research Systems. It took place in the Hackers and Designers Summer School 2015. A new series of workshops involving warm hearty foods is planned for February 2016. For more information visit www.algorithmickitchen.org