How To Build The Future: Max Hodak

A Neural Engineering Approach to Medicine

The core philosophy is shifting from a drug discovery paradigm to a neural engineering one. Humanity has historically struggled with drug discovery, often spending decades on research with no results. In contrast, we are empirically much better at engineering the brain. Brain-computer interfaces (BCIs) allow for a fundamental reframing of medicine by directly interfacing with the nervous system to restore or enhance function. This represents a paradigm shift in healthcare, moving beyond incremental biotech to transformative engineering solutions.

The Prima Retinal Implant: Restoring Form Vision

Science has developed a retinal prosthesis, the Prima implant, to restore vision to patients who have lost their rods and cones due to conditions like macular degeneration or retinitis pigmentosa.

• Mechanism: The system consists of a tiny 2mm x 2mm silicon chip, essentially an array of solar panels, implanted under the retina. The patient wears glasses equipped with a camera and a laser projector. The laser projects an image into the eye, and wherever the light hits the implant, it generates an electrical current that stimulates the retinal cells directly above it.

• Targeting the Right Layer: A key breakthrough was understanding which layer of retinal cells to stimulate. The retina has multiple layers of neurons. Previous attempts, like those by the company Second Sight, stimulated the retinal ganglion cells—the final output cells that form the optic nerve. This is after a significant amount of processing and compression (from ~150 million photoreceptors down to 1.5 million optic nerve cells) has already occurred. Stimulating at this stage results in patients seeing disconnected flashes of light, or “phosphines,” because the complex neural code for images is not being replicated.

• The Breakthrough: The Prima implant targets the bipolar cells, which are upstream from the ganglion cells and before this major compression step. The representation at this layer is much closer to a direct bitmap of the visual input. By stimulating these cells, the implant leverages the retina’s own processing capabilities. This was an empirical discovery from Science’s clinical trial and is the first time “form vision”—a coherent, assembled image in the mind’s eye—has been restored in this way. Patients in the trial, who had been unable to see faces for a decade, were able to read every letter on an eye chart.

The Brain as a Computer: APIs and Latent Spaces

The brain can be understood, almost literally, as a computer with a very different architecture. All information flows in and out of the brain through a well-defined interface, or “API”: the 12 cranial nerves and 31 spinal nerves. Reality, as we experience it, is constructed entirely from the electrical spikes traveling along these nerves.

This information processing perspective has seen a great unification with modern AI research.

• Neural Representations: As you move deeper into the brain from sensory inputs (like the retina) or outputs (like motor cortex), the neural representations become more abstract. For example, in the infratemporal cortex, there exists a “map of objects” where different clusters of neurons represent concepts like a vase, a car, or a face. Moving around in this high-dimensional space corresponds to perceiving different objects.

• Connection to AI: This is directly analogous to the latent space in AI models. Researchers have found that the internal representations learned by image and language models look remarkably similar to the neural representations found in the brain. This suggests that AI research is on the right track to understanding intelligence. A decade ago, the expectation was that AI would learn from neuroscience; today, the reverse is happening, with neuroscience learning a great deal from AI.

Neuroplasticity: The Adaptable Brain

The brain’s ability to adapt, or neuroplasticity, is a critical component of how BCIs work.

• Critical Periods and Adult Plasticity: While there are critical developmental periods that, if missed, can permanently affect brain wiring, the adult brain remains far more plastic than is commonly appreciated, especially when given direct feedback. The brain can quickly learn to control the firing of a single neuron if it can perceive the result.

• A Two-Way Street: BCI systems are not just about a computer decoding a fixed brain signal. The brain also adapts to the BCI. Early motor decoders worked by fixing the weights (e.g., “neuron A firing means move cursor up, neuron B means move right”) and simply letting the brain learn to control the output through feedback. This two-way learning system is incredibly powerful.

• The “Energy Well” Analogy: In adulthood, the brain’s plasticity isn’t always obvious because it has settled into a stable state—a deep “energy well”—after a lifetime of fitting itself to reality. It takes a significant new input, like a BCI, to reveal how adaptable it still is.

The Future: Bio-Hybrid Interfaces and Longevity

While current BCIs are focused on restoring function for disabled patients, future interfaces promise much more.

• Bio-hybrid Interfaces: Science is developing a novel interface inspired by nature. The idea is to create what is essentially a new cranial nerve with a “USB port on the end.” The device is seeded with living, lab-grown, hypoimmunogenic (hidden from the immune system) neurons. When this device is engrafted onto the brain, these new neurons physically grow and form biological connections with the host’s brain, creating an ultra-high bandwidth, bidirectional interface without penetrating electrodes. This is compared to the biological plugs seen in the movie Avatar.

• The Takeoff Era: We are entering a period of exponential technological change, a “takeoff era” comparable to the Industrial Revolution. The parallel advancements in AI and BCI are not yet fully appreciated. BCI is not just an AI-adjacent story about merging with machines; it’s a profound healthcare and longevity story. These technologies will reframe the human condition, with the potential for the first thousand-year-old humans to be alive today.

• Tags:
• AI
• Feature
• Neuroscience
• BCI
• Neuralink
• Retinal Implant
• Bio-Hybrid
• Neuroplasticity