CAULDRON REPORT

In early 2026, researchers at the Australian biotech startup Cortical Labs successfully trained a network of roughly 200,000 lab-grown human brain cells to play a simplified version of the classic 1993 first-person shooter, DOOM. 

Key Details of the Experiment:

• The Subject: The cells are living human neurons grown on a microelectrode array called the CL1 biological computer, which translates game data into electrical signals the cells can "feel".

• Goal-Directed Learning: Scientists used the Free Energy Principle to train the cells. When the neurons performed "correct" actions (like shooting an enemy), they received structured, predictable electrical feedback. Incorrect actions or "dying" resulted in chaotic, unpredictable noise, which the neurons naturally reorganized to avoid.

• Progression: This follows a famous 2022 experiment where a larger group of 800,000 neurons (nicknamed DishBrain) learned to play the 2D game Pong.

Do we really need to ask whether rapidly advancing humanoid robots deserve deeper scrutiny?

The issue is not technological progress itself. Innovation has always reshaped civilization — from the printing press to the internet. The real question is direction and intent. What are these systems being designed to replace, enhance, or control? And who ultimately benefits?

As humanoid robotics and neural-integration technologies accelerate, the conversation about capability is moving far faster than the conversation about sovereignty. That imbalance deserves serious attention.

The Psychological and Social Consequences

One of the most immediate concerns is social substitution.