Key Takeaways

  • Small-molecule drug discovery spends ten years and hundreds of millions of dollars before finding out if a biological target works.
  • Neural interfaces turn biology into an engineering problem with immediate feedback loops: placing electrodes into the motor cortex lets a user control a computer within an hour.
  • Science Corporation recently secured CE mark approval in Europe for PRIMA, a retinal implant that restores functional form vision to blind patients.
  • Because Science approaches biological dysfunction as hardware debugging rather than chemistry, the company raised most of its capital from Silicon Valley tech investors rather than traditional life science funds.

The Ten-Year Coin Flip in Drug Discovery

Traditional biotechnology runs on a brutal bet. A team spends a decade mapping molecular pathways, designing compounds, and running animal trials, all for a single clinical readout that often fails. As Max Hodak puts it: “You can do drug discovery for a decade, run a clinical trial, you're going to turn over a card, the answer might be no and then like what do you like you everybody goes home. Whereas here, we have a clear sense of how to make thing better.”

Human biology at the molecular level contains too many unmapped interactions. When a drug fails in phase 2 or phase 3 trials, researchers rarely get clear diagnostics on why the chemistry missed. You burn the capital, pack up the lab, and start from scratch on a new target. Humanity simply struggles with the sheer complexity of cellular chemistry.

Engineering Determinism in Neural Hardware

Treating the central nervous system as a computational circuit changes the iteration cycle completely. Circuits take inputs, run logic, and output signals. When you treat neural tissue as an electrical processor rather than a chemical soup, you can measure cause and effect on an oscilloscope in real time.

“The brain very literally, very clearly, plainly as a computer in my understanding of the world,” Hodak explains. “I also view the universe generally as a computer. Like, you can solve computational problems by arranging matter in a certain way and then taking your hands off and letting it go.”

This difference shows up in speed to signal. “If I put electrodes in M1, you will probably be using a computer in an hour,” Hodak notes. “And so it's just it's easier, it's more amenable to biology in many ways.” If the electrode array fails to register motor intent, engineers adjust electrode spacing, modify signal filtering, or update decoding algorithms that afternoon. The feedback loop drops from twelve years to twelve minutes.

This engineering predictability led Science Corporation to build PRIMA, a subretinal implant that replaces damaged photoreceptors to restore form vision. By translating optical data into electrical pulses that remaining retinal cells can process, PRIMA bypassed the molecular decay of macular degeneration entirely. The hardware earned CE mark approval in Europe because the physics produced clear, repeatable clinical outcomes.

Why Biotech Investors Missed the Boat

When you run a company with rapid hardware-software iteration cycles, traditional biotech metrics break down. Life science venture funds look for patent moats around molecules, target validation assays, and standard clinical phase milestones. They struggle to value companies that write firmware, design custom silicon, and debug device latency.

“Our device view of a lot of historical biology problems makes us even more of a tech company by biotech standards,” Hodak says. “So we mostly raised from tech investors, not that much from biotech investors.”

If your startup treats biological failure modes as information-processing bugs, do not pitch investors who only understand chemistry. Find partners who understand rapid prototyping, hardware tolerances, and software iteration speed.

What to Do With This

Audit your current product development bottlenecks this week. If you are solving an ambiguous problem with slow, high-cost experiments, ask whether you can reframe the problem into an engineering system with a direct feedback loop. Map the time delay between taking an action and observing the result. If your feedback loop takes months, redesign the testing surface to give you diagnostic data in hours before you burn more capital.