The Santa Ana winds were gusting strongly as the first worm simulation was initiated. Despite not being a hacker, the process was straightforward: opening a Terminal shell, pasting commands from GitHub, and watching the characters cascade down the screen, reminiscent of a scene from a movie. The code was scanned for familiar terms like "neuron" and "synapse" when a friend arrived to pick up the narrator for dinner. "One moment," was shouted from the office, explaining, "I’m just running a worm on my computer."
At a Korean restaurant, the atmosphere was frenzied; the winds bent palm trees and sent shopping carts skidding across the parking lot. It felt surreal, akin to a podcast played at double speed. A friend inquired, "You’re doing what, a cybercrime?" Amid the noise, an explanation was offered: "No, not a worm like Stuxnet. A worm like Richard Scarry."
Returning home in darkness, the first sparks had already reached Altadena. On the laptop, in a volumetric pixel box, awaited the worm. With pointed ends, it floated in a mist of particles, appearing eerily straight and still. Although not alive, it seemed lifeless. "Bravo," commented Stephen Larson later, acknowledging the achievement of reaching the "hello world" state of the simulation.
Stephen Larson, a cofounder of OpenWorm, has since 2011 spearheaded an open-source software initiative aimed at creating a computer simulation of the microscopic nematode Caenorhabditis elegans. The objective is to create a digital twin of the real worm with molecular accuracy. If successful, OpenWorm would be the first to produce a virtual animal and encapsulate extensive knowledge not only about C. elegans, a highly studied scientific subject, but also about brain-world interactions that result in behavior—deemed the "holy grail" of systems biology by OpenWorm.
Unfortunately, this goal has not yet been realized. The laptop simulation uses data from experiments on living worms, converting them into a computational framework called c302. This framework drives the simulated musculature of a C. elegans worm within a fluid dynamic environment—a representation of how the worm squiggles forward in a flat plate of goo. Generating five seconds of this behavior requires about 10 hours of computational time.
A lot can occur in 10 hours. An ember can drift on the wind from the foothills into the city. On Larson’s advice, the time parameters of the simulation were adjusted, moving past "hello world" deeper into the worm’s uncanny valley. The following morning, an unsettling orange haze loomed, and upon opening the laptop, two startling revelations emerged: Los Angeles was on fire, and the worm had moved.
At this juncture, a reasonable question arises, previously posed by the friend over banchan at the Korean restaurant: "Uhh … why?" Why, given the numerous challenges in the fragile global environment, would anyone invest 13 years in coding a microscopic worm into existence?
