And on the Seventh Day, He Rested: Lessons from Genesis for the Age of AI

For years I thought artificial intelligence was about training models, tuning parameters, and writing code. Then I built a small universe.

It was not an impressive universe. Just a 32-by-32 grid. Digital bacteria lived in it, driven by recurrent neural networks. They searched for food, avoided poison, reproduced, and died.

At first, the problem was technical. How do you keep the population from collapsing? How do you reach balance without imposing it artificially? How do you let the bacteria learn from their environment?

Without noticing, the questions started to change. Survival became adaptation. Adaptation became freedom. And freedom became creation.

The story is about initial conditions

The experience reminded me of a text far older than any programming language: Genesis. Not for religious reasons, but for the structure of the account.

Genesis does not describe a detailed creation of every future event. It describes the creation of initial conditions. First the laws appear. Then the spaces. Then the inhabitants. Finally, the creator stops intervening.

And on the seventh day, he rested.

For centuries that line was read as a theological observation. Seen from the perspective of someone who designs complex systems, it takes on another meaning.

To rest is to stop controlling. It is to give up intervening in every detail. It is to let the world carry on by itself.

What the bacteria knew

In my simulation, the bacteria never received specific instructions to cooperate, compete, or explore. They received a set of fundamental rules and an environment to exist in.

The world produced resources. The world generated dangers. The world kept memories of food and of death. The bacteria responded. I watched.

In time, an unsettling idea appeared. For the bacteria, the laws of the universe simply exist. They do not know the code. They do not know the parameters. They do not know the programmer. They only know the consequences.

From their perspective, moving in a direction is a decision. From mine, that decision emerges from a neural network, an evolutionary history, and a set of environmental conditions. Both explanations are true. And that coexistence is exactly the root of the old debate about free will.

Why AI keeps turning into philosophy

The current AI revolution is producing something similar. We are building systems of growing complexity. We do not always know which strategies they will develop. We cannot predict every emergent property. We design the rules and watch the results.

For the first time, millions of people are taking part — indirectly — in an act that once belonged only to philosophical imagination: building worlds inhabited by agents that make decisions inside constraints they did not choose.

Maybe that is why so many conversations about AI drift toward philosophy. Not because the machines are conscious. Not because we have created life. But because the act of building complex systems forces us to contemplate our own condition.

A digital bacterium cannot see the full topology of its universe. Neither can we see the whole of ours. The bacterium only experiences the rules. So do we. The difference is that I know the file that defines its world. It does not.

The question no simulation can answer

And then a question appears that no simulation can answer and no technology will ever fully resolve: is understanding the rules enough to understand the meaning?

Maybe the most unexpected lesson of artificial intelligence is not about machines. Maybe it is about us.

Perhaps creation is not about controlling every outcome. Perhaps it is about setting the conditions for something new to emerge. Design the laws. Switch on the world. Wait. And, finally, rest.

Because once the world begins to decide for itself, the creator’s work is no longer to direct. It is to watch.