C. elegans is a tiny worm, about one millimeter long. It has only 302 neurons.
And yet it can move toward food, avoid danger, learn from experience and change its behavior. Scientists have studied this worm for decades because its nervous system is simple enough to map. In 1986, researchers published the first complete wiring diagram of an animal nervous system. They could see which neurons connected to which.
It was a remarkable achievement. But knowing the wiring did not mean we understood the worm.
Even today, scientists are still working to understand how those 302 neurons produce behavior, memory and decision making. The connections matter, but so does the way signals move through them, change over time and respond to the environment.
I think about this when I look at what we are building for artificial intelligence. Modern AI uses artificial neural networks inspired by a simple idea from biology. Many small units are connected together. Signals pass between them. Through learning, the strength of those connections changes. We are now scaling this idea to an extraordinary size.
Around the world, companies are building enormous data centers filled with processors and connected by complex networks. The International Energy Agency estimates that electricity use by data centers could more than double by 2030, driven in large part by AI.
Perhaps this scale is necessary. We are asking AI to do things far beyond what a tiny worm needs to do. Still, C. elegans makes me wonder.
Nature has spent hundreds of millions of years learning how to do remarkable things with very little. A worm with 302 neurons can sense its environment, make choices and adapt. And after decades of studying its entire nervous system, we still do not fully understand how it works.
Maybe the future of AI is not only about building bigger machines. Maybe it is also about understanding how nature manages to do so much with so little. As we race to build larger data centers in pursuit of intelligence, a tiny worm offers a useful reminder. Sometimes the important question is not how many connections we can build. It is what those connections know how to do.
Daniel Kim, May 2025