Max Hodak on why BCI is no longer incremental, how his retinal implant restores form vision, and why the first people to live to a thousand are probably alive today.
The Takeoff Era
For decades, biotech has felt incremental—ten years of work for marginal gains. But something has shifted. Max Hodak, co-founder of Neuralink and founder of Science, says we are now firmly in a takeoff era. "Something new has happened on Earth," he says. The first people to live to a thousand, he believes, are probably alive today. That claim still requires suspension of disbelief, but the pace of change in brain-computer interfaces (BCI) is no longer incremental. Hodak sees the next fifteen years as transformative as the first decade of the steam engine—a period that would have been unimaginable to someone living just before the Industrial Revolution.
The key insight is that the brain is a computer. It processes information through a small number of cables: twelve cranial nerves and thirty-one spinal nerves. "That's the API of the brain," Hodak says. If you can read and write signals on those cables, you can restore lost function—sight, hearing, movement—and eventually create entirely new capabilities. The unification of AI and neuroscience has accelerated this: representations inside deep neural networks look remarkably like those in the brain, giving researchers a new way to understand neural codes.
Restoring Sight with Form Vision
Science's first product, Prima, is a retinal implant that has already given form vision to blind patients. The device is a tiny 2mm x 2mm silicon chip implanted under the retina, covered with microscopic solar panels. Patients wear glasses with a camera and a laser projector that beams an image into the eye. The chip absorbs the laser light and excites the bipolar cells above it, bypassing the dead rods and cones. In a large European clinical trial published in the New England Journal of Medicine, patients who had been unable to see faces for a decade could read every letter on an eye chart.
This was not obvious from the start. Earlier attempts, like Second Sight, stimulated the retinal ganglion cells—the optic nerve cells—and produced only flashes of light, not coherent images. The retina compresses 150 million rods and cones into 1.5 million optic nerve fibers, and that compression layer (the bipolar cells) does critical computation. By stimulating the bipolar cells before that compression, Science preserves the natural image representation. "It was an empirical discovery," Hodak says. "If you excite the bipolar cells with an image, you get an image in the mind's eye." The implant is black-and-white and has a small field of view, but it is genuine vision. The company is now submitting for regulatory approval and starting a new trial for inherited retinal diseases.
The Brain's Plasticity and Its Limits
Neuroplasticity is often misunderstood. There are genuine critical periods in early development—patients born with congenital cataracts who had them fixed as adults could not make sense of visual information, and several committed suicide. But the adult brain remains far more plastic than commonly appreciated. If you put an electrode almost anywhere in the cortex and give the patient feedback, they can learn to control a single neuron's firing within minutes. "The brain is very plastic under feedback," Hodak says.
Why does adult plasticity seem limited? The brain has settled into a stable attractor—a deep energy well shaped by years of fitting to reality. Psychedelics may temporarily shrink that landscape, but the brain quickly descends back into its familiar basin. This stability was selected for: it keeps perception reliable. But it also means that learning new representations requires overcoming that inertia. For BCI, this is both a challenge and an opportunity: the brain can adapt to new inputs, but the interface must be designed to work with the brain's existing representations.
Biohybrid Interfaces: Growing a New Cranial Nerve
Beyond electrical implants, Science is developing biohybrid neural interfaces—devices seeded with living, engineered neurons. The idea comes from nature: the corpus callosum connects the two brain hemispheres with 200 million fibers, creating a unified conscious experience. What if you could grow a new cranial nerve, like an "internet nerve," with a USB port at the end? Hodak's team cultures hypoimmunogenic stem-cell-derived neurons onto a chip, then grafts it onto the brain. The neurons grow and wire into the host tissue, forming biological connections without any wires penetrating the brain.
This approach avoids the need for genetic modification of the patient's own neurons—a one-way door that can go wrong. If the graft cells die, the patient is no worse off. In animal models, the grafted neurons have spread throughout the brain. Hodak compares it to the ponytails in James Cameron's Avatar: a new cranial nerve with a connector. This is not yet in humans, but it represents a path to ultra-high-bandwidth brain-to-brain or brain-to-machine links. "I think that BCI is going to come to be seen as a category, like pharma," Hodak says. "There will be many BCI companies going after different applications with different probes."
From Neuralink to Science: The Startup PhD
Hodak's journey began in software, but he was always drawn to the brain. After college, he started a robotic cloud laboratory company, Transcriptic, which was a slog. "I strongly identified with Ben Horowitz's essay 'The Struggle,'" he says. When Sam Altman emailed him in 2016 about a brain-computer interface company, Hodak initially recommended others, then realized he should pursue it himself. He joined Elon Musk to co-found Neuralink, where he learned how to execute a technically complex company requiring multidisciplinary teams and infrastructure.
The key lesson: running a startup is an oral tradition. Most successful founders learn from a handful of Silicon Valley cultures. Hodak advises young builders to first figure out what they want, then be high-agency and persistent—but also to consider working for someone like Musk early on. "That dramatically leveled up my ability to do this and know how the game is played," he says. At Science, he is applying those lessons to three parallel projects: retinal prostheses, biohybrid neural interfaces, and a perfusion program (Vessel) that aims to make heart-lung machines portable and affordable, so patients like the 17-year-old on ECMO can live at home instead of dying when the filter clots.
The Event Horizon at 2035
Hodak admits he can no longer see past 2035. The twin plot lines of BCI and AI are accelerating faster than most people realize. "Intelligence is going to become widely available for those who have the agency to deploy it," he says. He is generally optimistic—his p(doom) is well below 50%—but he acknowledges that the implications are staggering. If a brain-computer interface is also a brain-to-brain interface, it opens territory that is hard to imagine from today's perspective.
He does not predict a cure for all disease by 2035, but he expects new lateral options that reframe the human condition. The first people to live to a thousand are alive now, and they may be more numerous than we think. "Earth as a whole is at a moment of exceptional change," he says. For Hodak, the work is about giving people back what they have lost—sight, hearing, movement—and then, eventually, giving them capabilities that no human has ever had. The takeoff era has begun.