Max Hodak co-founded Neuralink with Elon Musk. Then he left to start something quieter and, in many ways, more ambitious. His company, Science Corp, has built a tiny retinal implant that can restore functional vision to people who have been blind for years. It is not science fiction. It is already in clinical trials, and it is working.
The device is smaller than a grain of rice. It sits behind the eye and sends precise electrical pulses to retinal cells that have stopped responding to light. The AI system driving the implant translates visual information from a camera into patterns of stimulation that the brain can interpret as sight.
How the Implant Works
The system has three components. A pair of glasses equipped with a small camera captures what the wearer is looking at. A processor, about the size of a coin, translates that visual input into electrical signals. And the implant itself delivers those signals directly to the remaining healthy cells in the retina.
What makes this different from earlier retinal prostheses is the AI. Previous devices used fixed grids of electrodes that produced blurry, low-resolution phosphenes. Science's system uses machine learning to adapt the stimulation pattern in real time, based on what the patient is actually looking at. The AI learns each patient's unique retinal map and adjusts the signals accordingly.
The result is not perfect vision. Patients do not suddenly see the way they did before. But they can read large print. They can recognize faces. They can navigate a room without a cane. For someone who has lived in darkness for a decade, that is transformative.
Forty Patients and Counting
More than 40 patients have received the implant in clinical trials. One of them, a woman in her sixties who lost her sight to retinitis pigmentosa over fifteen years ago, recently read a full novel for the first time in over a decade. She described the experience as "seeing the world come back in pieces, then in waves."
The trials have shown strong safety profiles. The implant procedure takes about two hours and is minimally invasive compared to traditional brain-computer interfaces that require opening the skull. Because the device interfaces with the eye rather than the brain directly, the risks of infection and tissue damage are significantly lower.
The Bigger Picture for Brain-Computer Interfaces
Hodak sees the retinal implant as the first step in a broader platform. The same underlying technology, using AI to translate between digital signals and biological neural tissue, could eventually help with hearing loss, chronic pain, and neurological conditions like Parkinson's disease.
But unlike the more futuristic visions of brain-computer interfaces that promise to let healthy people merge with AI, Science is focused on medical need. The company's first priority is restoring function that people have lost. The goal is not enhancement. It is recovery.
The World Health Organization estimates that 43 million people globally are blind and another 295 million have moderate to severe visual impairment. Many of these cases involve retinal degeneration, exactly the condition that Science's implant is designed to treat. If the technology scales, it could become one of the most impactful medical applications of AI in history.
Why This Matters
There is a version of the AI story that is about chatbots and code generation and productivity tools. And then there is this version. A tiny chip behind someone's eye, guided by machine learning, giving them back the ability to see their grandchildren's faces.
Science Corp is a Y Combinator alum. It was founded by someone who helped build one of the most talked-about companies in tech history. But the work itself is not flashy. It is painstaking, careful medical engineering. And it is changing lives in the most literal sense possible.
Sources: Y Combinator This Week at YC newsletter (March 15, 2026), Science Corp clinical trial disclosures, World Health Organization global vision impairment data (2024), Nature Biomedical Engineering.
