Software Engineer Uses Randomization to Break Out of Predictable Daily Routines

Max Hawkins began to feel trapped by the highly optimized structure of his daily life in San Francisco. , bought a single-origin pour-over coffee from the highest-rated café according to Yelp, and biked exactly 15 minutes and 37 seconds along the optimal route to his job as a software engineer at Google.

After eight hours at work he would meet friends for a beer at a craft brewery or spend time in Mission Dolores Park. One afternoon at the office, while reading an academic paper, Hawkins encountered a study that tracked the movements of 100,000 anonymized mobile-phone users over six months.

The research found that human mobility tends to follow simple, repeatable patterns. " That realization prompted Hawkins to question how the structure of daily routines shapes life outcomes. The following Friday, while planning to meet a friend at a newly opened bar that matched his usual preferences for good beer, soft lighting and nostalgic indie music, he could not stop thinking about the mobility study.

He decided to test an idea he had long considered: introducing randomness into his decisions. Hawkins created a program that summoned an Uber to an unknown destination chosen only by the driver. The first trip delivered him and his friend to the emergency room at San Francisco General Hospital.

Continued use of the random-ride generator took him to locations in the city he had never visited, including a leather bar in the Castro district, San Francisco State University’s planetarium and a bowling alley on an unfamiliar side of town. He expanded the approach by building additional apps that randomized restaurant choices, music playlists and even tattoos.

Humans have used tools such as coins, dice and drawing sticks for centuries to outsource choices to chance. A 2019 study by researchers at the University of Basel in Switzerland found that participants who flipped a coin often followed its guidance or used their emotional reaction to the outcome to clarify their true preferences.

Michel Dugas, who developed a scale in the 1990s to measure intolerance of uncertainty, said people who struggle with uncertainty tend to seek excessive information or act impulsively. Dugas views random decision-making not as evidence of strong tolerance for uncertainty but as another form of avoidance.

The concept also relates to the explore-exploit trade-off in computer science. Algorithms that rely too heavily on exploitation, or recommending familiar options, risk boring users and failing to adapt. Exploration introduces uncertainty but allows systems to learn new preferences.

Research cited in the article indicates that exposure to low-risk novel situations can help build tolerance for uncertainty.

2015 Hawkins left his position at Google and committed fully to randomized living. He wrote an algorithm that selected places to live within his budget, planning to stay one to two months in each location. His first destination was Ho Chi Minh City, Vietnam, reached on a one-way ticket.

He later attended randomized events, including 14 gatherings in a single day in Berlin that ranged from a baby-photography meetup to a honey-sauna gathering. After several years of nomadic travel, Hawkins returned to the United States and continued the experiments with a partner.

They allowed an algorithm to choose driving destinations across the country. The process eventually brought them to Williamston, a rural town in North Carolina’s Inner Banks region that had once housed a prisoner-of-war camp during World War II and hosted freedom rallies in 1963.

While walking the town’s historic streets with no local connections or planned activities, Hawkins questioned the value of his approach. " Hawkins concluded there is no universal optimal balance between exploration and exploitation. The balance varies by individual and changes with time and circumstances.

The article notes that exploration can help people discover new preferences and develop greater comfort with uncertainty through low-risk experiences.