Who will taste the cherries?

In 2015, a design lead visited a Silicon Valley test kitchen where chefs and researchers were developing a smart oven capable of cooking Michelin-grade meals automatically. The oven used sensors to measure food properties like moisture, color, and temperature, while chefs filled out sensory forms to record subjective qualities such as taste and texture. The goal was to create an oven that could replicate expert cooking techniques without human input. However, the chefs struggled to translate their intuitive methods into data the oven could understand. For example, one chef tried to cook lamb under a broiler to mimic the effect of 'smacking it with a propane torch,' a technique the oven’s logic could not process. The oven relied on quantifying taste and food qualities, assuming these could be reduced to measurable data points.

Computism is the belief that the universe and everything in it, including human experiences, can be broken down into discrete, measurable data points. This ideology extends into technology, economics, and politics, suggesting that the world can be governed by code. The article traces this idea back to ancient Greek philosopher Democritus, who proposed that the universe is made of indivisible bits called atomos, and later to René Descartes, who modeled the human body as a machine. In the 18th century, Jacques de Vaucanson’s mechanical duck, which simulated digestion, was an early attempt to mechanize biological processes. The smart kitchen project exemplified computism by trying to quantify taste, an inherently subjective experience, into data that an oven could process.

The article highlights the tension between human subjectivity and the computist belief that experiences like taste can be reduced to data. Taste is contextual, personal, and shaped by individual perception, making it difficult to quantify. The philosopher Albert Borgmann described this process as turning experiences into devices—systems that deliver outcomes with minimal human input, abstracting away the human element. For example, the smart oven’s designers viewed manual adjustments by users as a 'liability' because it introduced unpredictability. The oven could not understand subjective terms like 'smacking' a food with a propane torch, illustrating the limits of data in capturing embodied experiences. Despite these challenges, the project assumed that taste and food qualities could be fully represented through data.

In the 2010s, design thinking—a repeatable problem-solving process—became popular in tech, particularly at firms like IDEO. This approach promised objectivity by using data to guide design decisions, such as in user experience (UX) design. UX design focuses on how users interact with interfaces, often through measurable outcomes like engagement, conversions, and efficiency. Over time, UX design shifted from a craft centered on human subjectivity to one focused on mechanical objectivity, where interfaces were optimized for quantifiable performance. For instance, Google tested 41 shades of blue for search result links to determine which performed best, reducing design decisions to data-driven choices. This approach abstracted the human element of design, prioritizing measurable outcomes over subjective judgment.

Modern interfaces, such as those in smart kitchens or social media platforms, often function as mechanisms for extracting data from users. Techniques like A/B testing, where different versions of an interface are compared to see which performs better, became standard in tech. These methods treat users as data-generating objects, with their actions reduced to computable outcomes. For example, an endless feed with a pull-to-refresh feature exploits the user’s physiology to maximize data generation, encouraging them to stay engaged with the platform. The philosopher Byung-Chul Han argues that this shift reflects a broader move from physical coercion to psychological optimization, where users are driven to generate data willingly. The interface, in this context, becomes an autonomous subject, while the user is reduced to an object supplying data.