From Restoring Sight to Reimagining the Brain, with Max Hodak

Summary

• Science just got CE marketing approval in Europe (July) for Prima, its retinal prosthesis, with first commercial sales “in the coming weeks” — Hodak says it is the first to restore a form-vision image to a blind patient in this way. Trial patients with macular degeneration were filling in Sudoku and crossword puzzles and reading books; Hodak calls it “a great proof of concept that we’re on the right track,” with an engineering roadmap to add grayscale depth and “at least red and green” (blue is trickier).
• The deal engine behind Prima: Science found French company Pixium’s Stanford-originated implant to be “by far the state-of-the-art” in late 2022, then got to know the company for about two years before acquiring it. Hodak says the pipeline — vision, biohybrid neural interfaces, and the Vessel perfusion program — is the “minimum set” that could drive a significant revolution in medicine on a 10–15-year timescale if successful.
• Hodak’s core thesis is that “the brain very literally, very clearly, plainly is a computer,” and that treating it as one can produce unusually large effect sizes that are uncommon in medicine. Small-molecule discovery means a decade of work to “turn over a card” where “the answer might be no, and then everybody goes home” — versus “if I put electrodes in M1, you’ll probably be using a computer in an hour.”
• Hodak does not see a “brain keyboard” as Science’s focus: “talking or writing is thinking,” and a deeply evolved ~10-bits-per-second cognitive bottleneck rolls up through language. The real prize for Science is on the other end of the spectrum: “if you get vision, hearing, balance, and a kilobit per second of motor control, you’re halfway to the Matrix.”
• The Platonic Representation Hypothesis — that AI models and brains may share underlying representations — is used “constructively” at Science, which gets alignments between animal neural recordings and AI model internals. “It feels like a law of physics. If you apply enough compute to matter, you get this thing that looks like intelligence” — a clue for Hodak that AI was “not a gimmick and not hitting a wall.”
• Pricing isn’t set, but precedents are rich: Second Sight got ~$150,000 per patient in the mid-2010s for mere flashes of light, and a gene therapy with about a 0.1-line improvement reimburses at almost $500,000 per eye. Hodak says “like 1 in 2” have some early-stage AMD by 80 and “like 1 in 10 at 85” actually have it; he puts the first-version US/Europe population at hundreds of thousands, while the next version, in animal studies now and hopefully in humans next year, should expand that to millions.
• The 20-year vision is substrate independence and reduced human fragility: “that sense of jeopardy will fade” as parts become upgradeable and swappable. Cardiovascular disease and brain-metastasizing cancer look “really attackable”; neurodegeneration “still seems difficult” — and “I’m gonna be ultimately fairly disappointed if I’m murdered by my pancreas.”

Deep dive

1. Prima is approved — and “people forget that the moonshot worked”

• The product: a tiny chip implanted under the retina of patients who’ve gone blind from loss of light-sensitive cells; glasses with a laser projector beam an image onto the implant, which stimulates the retina directly, bypassing dead rods and cones. Trials were in macular degeneration; studies in retinitis pigmentosa and Stargardt disease are planned. CE marketing approval landed in July after about two years of post-acquisition work, with first sales due in the coming weeks.
• The path there: Science considered the retina, the LGN in the thalamus, and V1 (“half a billion cells”) as entry points for vision; it also explored electrical stimulation, ultrasound, and gene therapy. It built an in-house gene therapy (“probably going into humans next year”) and later acquired Pixium, a French company whose Stanford-originated retinal-stimulation work was, in Hodak’s view, “by far the state-of-the-art” in late 2022. Science got to know Pixium over about two years before acquiring it. “That deal has turned out to be great.”
• Guo’s framing — that BCI is seen as a moonshot — draws a correction: “we went and left boot prints on the moon.” Silicon Valley has repurposed the word to mean long odds “and therefore we can vaporize a bunch of investor money just fine”; Hodak says moonshots have historically succeeded more than people give them credit for.
• The trial’s main output was the existence proof: patients doing Sudoku, crosswords, and reading books — “one of those things that seems too good to be true.” Current limits: “it’s like looking through a straw,” black and white only, with an engineering path to greater grayscale depth and at least red and green; blue is trickier.

2. “If you wanna make people angry, tell the internet that the brain is a computer”

• Hodak means it literally, not metaphorically: you solve computational problems “by arranging matter in a certain way and then taking your hands off and pressing go.” The universe is a computer too; “there’s nothing special about transistors.”
• Science’s three-part pipeline is vision, biohybrid neural interfaces — grafting living neurons rather than placing metal wires or genetically modifying the brain — and Vessel, its perfusion program. Hodak calls these the minimum set that could drive a broad medical revolution over 10–15 years if successful.
• The tradeoff in his framing: humanity “just isn’t that good” at the biological understanding required for drug discovery — decade-long discovery, a clinical trial, “you’re gonna turn over a card, the answer might be no, and then everybody goes home” — while devices offer a clearer path to incremental engineering improvements. Even highly engineered CAR-T can produce “a giant immune overreaction,” but a quadriplegic with motor-cortex electrodes can be “playing video games in, like, an hour.”
• Positioning follows culture: Science sits on the tech side of biotech’s “East Coast–West Coast divide,” raised mostly from tech investors — the only biotech VC Hodak sought at the Series A was Bob Nelson.

3. BCI is a category like pharma — and brain keyboards are not Science’s focus

• Hodak’s jab at investors: “I’ll talk to VCs, and they’re like, ‘Oh, we have a BCI bet.’ I’m like, ‘Do you have a drug bet? You made one bet out of drugs.’” Silent-speech and similar neural-input devices may be BCI to a greater or lesser degree, but he sees them as basically hand substitutes. Hands already work well, and he argues that products controlling apps would need unambiguous intent rather than accidentally summoning two Ubers mid-meeting.
• His skepticism about latent-thought decoding, in response to Guo’s question about a “special latent state that’s not language”: the feeling that thoughts are preformed “is misleading” — you find out when you sit down to write. The ~10-bits-per-second bottleneck (a perfect-memory subject drawing Manhattan after a helicopter ride works out to about that rate) means brain keyboards might be like AR glasses: attention “was already fully 100% occupied.”
• The frontier he does care about is the transition from communicating with a thing to “redrawing the border around your brain.” Language works by lighting up “pre-shared concept spaces” between brains; the hemispheres of your own brain are bound differently — “figuring out where that transition happens is a really compelling area of research.”

4. Substrate independence turns on continuity, not copies

• The thought experiment: scan yourself non-destructively, let the software replica “keep doing my venture-investing job” while you go to hospice — “does that make you feel better, that much?” Guo raises general anesthesia; Hodak says it produces a break, that people still find it different from a copy, and that this difference needs explaining.
• His camp: “continuity is greatly important” — people accept significant identity drift with continuity, but a replica that answers exactly like you yet isn’t phenomenally continuous is “less satisfying.” Guo says she would trade dramatic morph for continuous experience but is unsure about degraded IQ (a question Laura Deming asked her); Hodak suggests temporary degradation might be acceptable if it were restored weeks later — “where you achieve substrate independence, you can take that almost anywhere you want.”
• The missing piece is connectomics: a human connectome is still relatively far away, but “we’re not that far from a mouse connectome,” which Hodak says would be enormously useful for understanding the brain, including basic questions such as its overall architecture.

5. AI models and brains may grab the same manifold — so do neuroscience on the models

• The Platonic Representation Hypothesis: the mathematical objects inside big AI models “look a lot like the things that you see in neuroscience,” with “very similar geometry” for concept representation. Science uses this constructively, aligning animal-brain recordings with AI model internal representations — Hodak’s early clue that AI “was on the right track.”
• His strongest claim remains hedged as an intuition: “there’s some fact about the universe where these things as they’re learning are grabbing onto some true underlying data manifold… it feels like a law of physics.”
• Why it’s controversial: “some faction of people that kind of don’t want this to be true for reasons that are not totally clear to me,” plus genuine unknowns — whether the structure is global or local. Asked for the most fertile way to study neuroscience, Hodak says “ironically, it’s probably working on AI.” He jokes with neuroscience friends at OpenAI and Anthropic that they “left neuroscience”; his point is that it is easier to do neuroscience on models.

6. The business case, and why the other organs are “support characters”

• Pricing is deliberately unannounced, but the comps: Second Sight got ~$150,000 per patient in the mid-2010s without form vision; a gene therapy relevant to ~5% of patients in one narrow indication, delivering about a 0.1-line improvement and slowing degeneration for some patients, reimburses at almost half a million dollars per eye. Hodak says “like 1 in 2” have some early-stage AMD by 80 and “like 1 in 10 at 85” actually have it; he estimates the first version’s US/Europe population at hundreds of thousands, the current version of Freedom at probably hundreds of thousands, and the next version at millions.
• The worldview underneath: the brain is the only organ you can’t even in principle transplant; heart, pancreas, liver, and lungs “are really support characters.” Biology is “alien nanotechnology… that we understand still very poorly” — so route around it with toolboxes humanity is more advanced in. “I’m gonna be ultimately fairly disappointed if I’m murdered by my pancreas.”
• The 20-year picture: the “sense of jeopardy” of the human condition fades, with cardiovascular disease and cancer that metastasizes to the brain — which Hodak calls the two leading causes of death — “really attackable,” while neurodegeneration “still needs real investment.” Adapting ourselves to “the hard vacuum of space” and achieving substrate independence “are the same project.”