We've all heard the unwritten rule of technology: if it exists in nature or invention, someone will eventually run Doom on it. From calculators and pregnancy tests to tractors and even gut bacteria, the 1993 classic has been ported to the most unlikely platforms. But now, Australian biotech firm Cortical Labs has escalated things to pure sci-fi territory by teaching a cluster of living human brain cells to navigate the game's demonic mazes and blast away at enemies.
.jpg)
This isn't just a gimmick — it's a leap in biological computing. Back in 2022, the same team made headlines when their "DishBrain" system, consisting of about 800,000 neurons grown on a silicon chip, learned to play the simple 2D game Pong.
In that experiment, the neurons were stimulated with electrical signals representing the ball's position, and their responses moved a virtual paddle. The cells adapted quickly, preferring predictable stimuli over chaotic ones, effectively learning to extend rallies to avoid "punishment" in the form of random electrical noise.
Fast-forward to February 2026, and Cortical Labs has upgraded to their CL1 biological computer, a rack-mountable device that houses around 200,000 human neurons derived from stem cells. These neurons are suspended in a nutrient-rich liquid on a multi-electrode array, allowing the chip to send and receive electrical signals in real time. The breakthrough? An independent developer, Sean Cole, used a Python-based API to interface the neurons with Doom in just a week.
Here's how it works: The complex 3D world of Doom — complete with corridors, enemies like cacodemons, and actions like moving, turning, and shooting—is translated into the only language neurons understand: electricity. Game data, such as enemy positions or ray-cast distances from walls, is encoded into patterns of electrical stimulation across the chip's electrodes.
When an enemy appears on the left, for instance, the corresponding neural region gets zapped. The neurons fire back with impulses, which a decoder interprets as commands: fire a shot if they spike in one pattern, strafe right in another. Essentially, the brain cells are trying to minimize discomfort by adapting their responses to make the "shocks" more predictable, much like avoiding pain in real life.
The results are impressive yet hilariously humble. The neural network outperforms a random player but lags far behind even novice humans—think running into walls and shooting at ceilings, like your dad fumbling with his first mouse. It learned faster than traditional AI systems, mastering basic gameplay in days rather than requiring massive datasets. However, limitations abound: the neurons don't "see" the game visually; they process abstract electrical inputs. Performance is inconsistent, cells die frequently, and the system isn't truly conscious — it's more like a biological filter that influences decisions.
.jpg)
Cortical Labs' CTO, David Hogan, explained in a demo video that this setup demonstrates adaptive learning in biological systems. The company sells the $35,000 CL1 for research, accessible via their cloud platform, and envisions applications beyond gaming: controlling robotic arms, modeling brain diseases, testing drugs, or creating energy-efficient computers that learn like living organisms. Human brains run on just 20 watts, after all, compared to power-hungry AI servers.
This experiment has sparked buzz online, with X users calling it "insane" and pondering its implications for hybrid AI. As one post quipped, these cells played Doom before GTA 6 dropped. While it's not esports-ready, it's a step toward merging biology and silicon, blurring the lines between life and machine in true cyberpunk fashion. Who knows? Next up might be neurons tackling modern titles — or solving real-world problems.
Also read:
- The Invisible Blockade: How Insurance Shut Down the World's Energy Lifeline
- How MrBeast's Team Skyrocketed to +117 Million Subscribers in 2025: The Power of Aggressive Google Ads
- Anthropic Introduces Memory Import in Claude: Bring Your Context from Any AI Tool
Thank you!
