You know the internet’s favorite benchmark for any piece of hardware, right? From graphing calculators to smart fridges, the ultimate question for tech enthusiasts has always been: "But can it run Doom?" Well, we finally have an answer that bridges the gap between retro gaming and hardcore science fiction. Yes, a literal clump of human brain cells can play Doom.
In a move that feels like the prequel to a cyberpunk novel, Australian startup Cortical Labs has successfully taught a cluster of approximately 200,000 living human neurons to navigate the corridors of the classic shooter (specifically the open-source version, Freedoom). This isn’t a simulation of a brain; it is actual biological matter housed in a system called the CL1.
This development marks a massive leap forward from 2022, when the company’s "DishBrain" experiment taught neurons to play the 2D classic Pong. But navigating a 3D environment while managing multiple inputs? That is a whole new level of complexity for what the industry is calling "Organoid Intelligence" (OI).
How does the CL1 biological computer actually work?
So, how exactly does a petri dish learn to strafe? The CL1 is what researchers call a "biological computer." It houses the neurons on a high-density microelectrode array that interfaces directly with the living cells. Think of it as a translator that turns digital information into biological signals.
The learning process is fascinatingly simple in concept but complex in execution. The system uses "goal-directed learning." The game’s video feed is converted into electrical signals that the neurons receive. The neurons are programmed—in a biological sense—to minimize chaotic feedback. In the context of the game, playing properly produces a predictable, rhythmic pattern of stimulation. Dying or hitting a wall creates chaotic, unpredictable feedback.
Because biological neurons naturally want to minimize that "noise" or prediction error, they physically reorganize themselves to fire in patterns that keep the game going. They aren’t playing for a high score; they are playing to avoid the annoyance of chaotic electrical signals.
Is the biological neural network good at playing video games?
Before you start worrying about a jar of brain cells beating you in a multiplayer match, let’s manage expectations. Dr. Brett Kagan, the Chief Scientific Officer at Cortical Labs, is pretty transparent about the current skill level. He notes that the cells are absolutely not eSports champions.
Right now, the performance is described as "stumbling." The neurons behave like a beginner who has never seen a computer before. They can locate enemies, move around, and even shoot, but they die frequently. However, the fact that they can interpret a 3D environment and react to it at all is the breakthrough here. It proves that wetware computing can handle dynamic, multi-variable environments, not just simple paddle-and-ball physics.
How much does a wetware computer cost?
This is where the story shifts from cool science experiment to actual product. The CL1 isn’t just staying in a university lab. Cortical Labs describes it as the "world’s first code deployable biological computer," and they are putting a price tag on it.
If you are a researcher or a very well-funded hobbyist, you can buy a CL1 unit for approximately $35,000 USD. Commercial shipments, initially scheduled for June 2025, have already commenced. If that upfront cost is too steep, the company is also offering a rental model—"Wetware-as-a-Service," if you will—for about $300 a week. This allows scientists to access the biological compute power remotely without maintaining the cells themselves.
They aren’t the only ones in this space, either. Swiss competitor FinalSpark recently launched a "Neuroplatform" that offers remote access to biological networks for around $500 a month, and other players like Koniku are exploring similar tech for biological sensing. The market for this tech is projected to jump significantly by 2034, according to various market research reports.
The Bottom Line
While watching brain cells play Doom is an incredible headline, the real story here is energy efficiency. Silicon chips are approaching their physical limits and are notoriously power-hungry, whereas biological brains are the most efficient learning machines in the universe. Cortical Labs isn’t trying to build a better gamer; they are building a compute architecture that could eventually replace silicon for complex, low-power AI tasks. The winners here will initially be researchers in drug discovery who can test toxicity on actual human neurons without animal testing, but the long-term implication is a fundamental shift in how we build computers.
![Illustration related to Human Neurons Playing Doom: How the CL1 Works [Explained]](https://bytewire.press/wp-content/uploads/bytewire-images/2026/03/human-neurons-playing-doom-cl1-explained-ea564de82b.webp)
![Diagram related to Human Neurons Playing Doom: How the CL1 Works [Explained]](https://bytewire.press/wp-content/uploads/bytewire-images/2026/03/human-neurons-playing-doom-cl1-explained-2502f7e8e8.webp)
