Google has created a detailed map of the brain and central nervous system of a fruit fly, known as the MaleCNS v1.0 dataset. This dataset includes more than 166,000 neurons from an adult male fruit fly, developed through a multi-year collaboration between the HHMI Janelia Research Campus, Google Research, and other partners. Using artificial intelligence, researchers converted millions of 2D images into 3D reconstructions of the fly’s neural structures. This work marks a significant step in neuroscience, offering a comprehensive biological map that can be used for further study and research. The dataset has sparked creative and unexpected applications. For example, software engineer Alex Wormuth connected the fruit fly’s neural map to the video game Doom. In this setup, each image from the game stimulated the sensory neurons of the simulated fly brain. Neural activity was linked to game commands, and damage in the game triggered a stimulus toward specific brain cells, mimicking a form of reinforcement learning. Wormuth made his code open source and created a live page to show how the model, made up of around 166,700 neurons, learns to survive in the game. He questioned whether the fly would eventually learn to navigate and survive in the game environment. A similar experiment was conducted by developer Jessica Paquette, who connected the fruit fly connectome to Super Mario 64. A short video she shared showed Mario repeatedly bumping into a wall, indicating the system is still in its early stages. Paquette described the project as a fun experiment, entirely coded by instinct using GPT Astra. The code for this project is also open source, allowing others to explore and build upon the work. These experiments highlight how a detailed biological map can be used by software developers to explore how neural networks respond to external inputs and feedback. While the connectome was initially created as a scientific resource, it has also become a public testing ground for developers. This approach allows researchers to study how known neural connections behave in new and complex environments. In March 2026, scientists constructed a biological computer using human neurons grown in a laboratory. The system was taught to play Doom using electrical stimulations and a feedback system. The long-term goal was to understand how neurons learn and adapt, which could have implications for pharmaceutical research and computing. However, some scientists raised ethical concerns about "living robots," questioning their sensitivity, environmental impact, and moral status. They called for regulatory oversight. In December 2021, researchers at the biotechnology startup Cortical Labs demonstrated that brain cells grown in a petri dish could learn to play Pong faster than an AI. The "mini-brains," made up of 800,000 to one million living human brain cells, learned the game in just five minutes, compared to 90 minutes for an AI. This experiment showcased the potential of lab-grown neurons for learning tasks. In June 2025, FinalSpark, the company behind Neuroplatform, began offering paid remote access to its bioprocessors, allowing university researchers to use shared organoids for $500 per user per month. The Australian startup Cortical Labs also introduced the CL1, a computer that combines human stem cell-derived neurons with silicon. These developments have sparked discussions about both the technical capabilities and ethical implications of such technologies.