Australian startup Cortical Labs has unveiled the CL1, a commercial biocomputer that combines living human neurons with silicon chips. The development is capable of processing information in real time and potentially opens up a new era in brain research, drug development and AI.
Australian startup Cortical Labs has unveiled the CL1, a commercial biocomputer that combines living human neurons with silicon chips. The development is capable of processing information in real time and potentially opens up a new era in brain research, drug development and AI.
According to Spectrum, the world’s first biocomputing platform that can be uploaded to code was created by a company Cortical Labs. The CL1 device uses approximately 800,000 human neurons grown from reprogrammed skin or blood cells from adult donors. They are connected to silicon chips, forming a closed-loop system that can learn, adapt to stimuli, and interact with software via electrical signals with a latency of less than a millisecond.
The device is aimed at research institutes, biotech companies, and universities working in neuroscience, drug development, cognitive modeling, and even cryptocurrency mining. Each CL1 costs $35,000, or $20,000 when ordering a server rack of 30 units. In addition, Cortical Labs offers «wet iron as a service» — access to neural cultures in the cloud for $300 per week.
The platform has already shown successful recovery of function in artificial epileptic brain models after drug exposure. Meanwhile, neural networks in CL1 learned to play Pong faster than deep learning models, even during the DishBrain prototype stage.
The CL1 has 59 input channels (up from 8 in its predecessor) and a built-in cell life support system. The company plans to scale up, as biological systems allow for cheap cell growth to hundreds of millions.
The Melbourne-based company first gained attention in 2022 when its prototype learned to play the arcade game Pong. The project’s scientific inspiration is Carl Friston’s free energy minimization theory, which explains how the brain learns through prediction. Friston has called the CL1 a «brain in a jar,» opening up new horizons for experimenters in studying the function of neural networks.
Nvidia recently introduced its new compact supercomputer for expanding the capabilities of generative AI, which is already available for sale for $249. The unique Jetson Orin Nano Super Developer Kit fits in the palm of your hand.
