Mind meets silicon: when living brains take on video games and full simulations

In early 2026, Cortical Labs hooked up roughly 200,000 living human neurons on its CL1 chip and taught them to play DOOM. Building on the 2021 breakthrough where the same biological hardware mastered Pong, an independent developer used simple Python code to train the neurons in just one week. Electrical pulses fed the game’s visuals and threats straight to the cells, and their firing patterns moved the character and pulled the trigger. The petri dish held a real biological computer, and it learned a complex game faster than many expected.

At the same time, scientists made a similar advance with an insect. In October 2024, a global team completed the first full connectome of an adult fruit fly brain: every one of its 139,255 neurons and all 50 million synaptic connections mapped in 3D. Phil Shiu and colleagues at UC Berkeley then loaded that entire wiring diagram into a laptop simulation. The digital fly brain reacted to virtual tastes and touches exactly like a living one would, and the simulation predicted neuron activity accurately. It’s the largest complete brain ever simulated in software, and it runs on ordinary hardware.

These two advances, one wet and alive and the other fully digital, pull biology and computing together. Human neurons playing DOOM show that living cells can be programmed like silicon. The fruit-fly simulation shows a complete nervous system can be copied and made to think. They hint at a future where biological computers power AI and full brain uploads move from sci-fi to lab bench.

Sources

Human brain cells playing DOOM - New Scientist
IFLScience, and Cortical Labs official video. Fruit fly connectome and laptop simulation: Berkeley News and Nature papers (2024).