A near-perfect simulation of the human brain would have profound implications for humanity. It could offer a pathway for us to transcend the biological limitations that have constrained human potential, and enable unimaginable new forms of intelligence, creativity, and exploration. This represents the next phase in human evolution, freeing our cognition and memory from the limits of our organic structure.

Unfortunately, it’s also a long way off. The human brain contains on the order of one hundred billion neurons — interconnected by up to a quadrillion synapses. Reverse-engineering this vast network would require computational resources far exceeding what’s currently available. Scientists seeking a proof of concept for whole brain emulation have had to turn to simpler model organisms. And by far the simplest available brain — at just 300 neurons — belongs to the nematode Caenorhabditis elegans. 

Scientists have been working on the problem of simulating C. elegans in some form or another for over 25 years. So far, they’ve been met with little success. But with today’s technology, the task is finally possible, and — as I’ll argue — necessary. 

The biologist Sydney Brenner became interested in C. elegans as a model organism for developmental biology in the 1970s. Its simplicity and small size made it an ideal lab subject. In 1986, John C. White, a scientist in Brenner’s research group, produced a nearly complete map of the neural connections that make up the C. elegans brain — what scientists now call the connectome. As computers became more accessible, other scientists started building on Brenner’s work. Ernst Neibur and Paul Erdös kicked things off with a biophysical model of nematode locomotion in 1991. Two different teams (one at the University of Oregon and the other in Japan) published plans for building more ambitious models in the late 1990s. Both would have utilized White’s work on neural circuitry.  Unfortunately, neither got off the ground. 

In 2004, the Virtual C. elegans project at Hiroshima University got somewhat farther: they released two papers describing their model, which simulated the nematode’s motor control circuits. The simulated nematode could respond to virtual pokes on its head, but it didn’t do much else. And even this was, arguably, not a true simulation. Although the researchers had a map of the nematode’s neurons, they didn’t know their innate biophysical parameter — that is, the precise electrical characteristics of the connections between them. Instead, the researchers used machine learning to produce a set of values for each neuron that made their simulated nematode respond to a poke like a real one would. As a result, this approach was not entirely grounded in biological reality — a recurring theme that would surface in several future simulation attempts.

That is where things stood at the dawn of the 2010s. While work continued on simulating nematode locomotion, there was no progress on simulating a nematode’s brain — let alone a realistic one. Then, on January 1st, 2010, the engineer Giovanni Idili tweeted at the official account of the Whole Brain Catalogue, a project to consolidate data from mouse brains: “new year’s resolution: simulate the whole C.Elegans brain (302 neurons)!” U.C. San Diego neuroscience grad student Stephen Larson noticed the tweet and, by August, Larson was pitching the idea at conferences. By early 2011, Larson and Idili had put together a team to start work on what would become the OpenWorm project — the efforts of a decentralized group of academics with the goal of creating a complete, realistic, and open source model of C. elegans. 

This was a heady time to be interested in simulating extremely tiny brains. Over the next few years, OpenWorm published a series of papers and model updates. In 2013, they hosted their first conference in Paris and landed an optimistic story in The Atlantic (title: “Is This Virtual Worm The First Sign of the Singularity?”). Meanwhile, the researcher David Dalrymple was working on a parallel project at MIT, which he dubbed Nemaload. OpenWorm scientists largely used data from dead nematodes but Dalrymple wanted to use the then-new technique of optogenetics to study living specimens. Optogenetics allows scientists to control neurons and other cells with light. In this case, the technique could be used to collect data on how a nematode’s brain responds to different states by perturbing it thousands-upon-thousands of times. In a 2011 comment on LessWrong, Dalrymple wrote “I would be Extremely Surprised, for whatever that’s worth, if this is still an open problem in 2020.” 

It’s now 2025, and nematode simulation remains an open problem. Dalrymple abandoned Nemaload in 2012. OpenWorm still exists but has not made substantial progress over the past ten years towards creating a truly scientific whole brain simulation, due to a lack of available data. Occasionally, more modern (though still heavily assumption-based) simulations are published, including integrative models that strive to make fewer assumptions. We’re not quite back where we were in the 2010s: we have much