After Neuralink, Max Hodak is building something stranger

Six years ago, I asked Sam Altman at a StrictlyVC event in San Francisco how OpenAI, with its complicated corporate structure, would make money. He said that someday, he’d ask the AI. When everyone snickered, he added, “You can laugh. It’s all right. But it really is what I actually believe.”

He wasn’t kidding.

Sitting again in front of an audience, this time across Max Hodak, the co-founder and CEO of Science Corp., I can’t help but remember that moment with Altman. Pale-complexioned Hodak, wearing jeans and a black zip-up sweatshirt, looks more like he’s going to jump into a mosh pit than pitch a company valued at hundreds of millions of dollars. But he’s got a sly sense of humor that keeps the room engaged.

Hodak started programming when he was six, and as an undergraduate at Duke, he worked his way into the lab of Miguel Nicolelis, a pioneering neuroscientist who has since become publicly critical of commercial brain-computer interface ventures. In 2016, Hodak co-founded Neuralink with Elon Musk, serving as its president and essentially running day-to-day operations until 2021.

When I ask what he learned working alongside Musk, Hodak describes a specific pattern. “We got into lots of situations together where something would happen. In my mind, I’d have two diametrically opposed possible solutions, and I would bring them to him, and I’d be like, ‘Is it A or B?’ And he’d look at it and be like, ‘It’s definitely B,’ and the problem would never come back.”

After a few years of this, Hodak took what he’d learned and roped in three former Neuralink colleagues to launch Science Corp. about four years ago. Like Altman, Hodak describes his team’s improbable goal so placidly that I find myself believing that the limits of cognition are about to be overcome sooner than most of us realize. And that he’ll be among those who make it happen.

While I’ve been consumed with the AI data center craziness and the talent poaching wars, momentum has been building in the background.

According to World Economic Forum data, nearly 700 companies around the world have at least some ties to brain-computer interface (BCI) technology, including some tech giants. In addition to Neuralink, ​​Microsoft Research has run a dedicated BCI project for the last seven years. Apple partnered earlier this year with Synchron, backed by Bill Gates and Jeff Bezos, to create a protocol that lets BCIs control iPhones and iPads. Even Altman is reportedly helping to stand up a Neuralink rival.

And in August, China released its “Implementation Plan for Promoting Innovation and Development of the BCI Industry,” targeting core technological breakthroughs by 2027, and aiming to become the global leader by 2030.

Much of the neuroscience isn’t new. “A legitimate criticism of the BCI companies is that they aren’t doing new neuroscience,” Hodak said. “Decoding cursor control or robotic arm control from a human – people have been doing that for 30 years.”

What’s new, however, is the engineering. “The innovation at Neuralink is making [a device] small enough and low-power enough that you can fully implant it and close the skin, and have something that isn’t an infection risk. That genuinely was new.”

Indeed, Hodak admits we’re missing a lot of information about how the brain works to really build the products that he’s talking about. But unlike a lot of BCI companies that have to raise money, Science Corp. is figuring out ways to generate revenue. On a small scale, it makes tools that it then sells to other researchers — as Hodak puts it, “taking a $300,000 cart-sized recording system and turning it into a $2,000 handheld.”

The bigger unlock is getting something to market soon. A product that can help people and make money while the company quietly builds technology that it claims could reshape human consciousness itself.

That initial commercial “product” is a procedure called Prima. It’s exciting enough that Time magazine put the tech on its cover a few weeks ago: a computer chip smaller than a grain of rice that’s implanted directly in the retina. Combined with camera-equipped glasses and (for the time being) a two-pound battery, the tech restores vision to people with advanced macular degeneration. Not blurry, vague light perception, but “form vision.”

In completed clinical trials with 38 patients, Science Corp says 80% were able to read again, two letters at a time. “To my knowledge, this is the first time that restoration of the ability to fluently read has ever been definitively shown in blind patients,” says Hodak.

Science Corp. can only take so much credit. It acquired Prima from a French company called Pixium Vision last year, refined the technology, completed the trials Pixium had started, and submitted the results for approval in Europe. Hodak expects to launch the product next summer.

Prima is meanwhile still to be approved by U.S. regulators. Asked about the FDA, Hodak said, “We’re working with the FDA, although there’s some questions on exactly the timeline for that.”

Either way, at what he estimates will cost $200,000 per procedure at the outset, Science Corp. will become profitable if it manages to get just 50 patients per month.

The mind is in a dish of neurons

The next, more ambitious step is gene therapy. Specifically, optogenetic gene therapy, which means making neurons light-sensitive so they can be controlled with light instead of electrodes. It’s not a new idea, but Science Corp. thinks it’s figured out what everyone else hasn’t.

Here’s how Prima works: Your retina has three layers of cells. Photoreceptors (rods and cones) at the back capture light, and connect to bipolar cells, which connect to optic nerves that run to the brain. In macular degeneration, the photoreceptors die. Prima’s 400 electrodes stimulate the bipolar cells directly, bypassing the dead layer.

With gene therapy, the goal is to skip the electrodes entirely. Instead, you engineer the surviving cells using new proteins to respond to light.

“The eye is a really ideal place to do this type of gene therapy work, because it’s kind of left alone by the immune system,” Hodak explains. In other parts of the body, engineered cells that express unfamiliar proteins trigger immune attacks. But our body learned long ago not to overreact when things change in the eye.

Other companies are pursuing similar approaches, but Hodak says they are either targeting the wrong cell layer, or their proteins just aren’t as good as his startup’s. “They’re not as fast, they’re not as sensitive. The proteins that we’re actually using are state of the art,” he claims.

Either way, even gene therapy isn’t the long game. That is something Hodak has probably been dreaming about his entire life: a way to grow new brain tissue.

Electrodes are crude; they damage tissue. “Every time you place something mechanically into the brain, there’s no free space in there,” Hodak explains. He says there’s redundancy in the cortex, and for someone with a spinal cord injury or blindness, the tradeoff is “totally justifiable.” But the tissue damage means “you can’t scale it up to millions or billions of channels.” That, he says, is the fundamental limitation of approaches like Neuralink’s.

Adding more neurons to the brain sounds batty, but Science Corp. says it has already tested a proof-of-concept device in mice. The device looks like a tiny waffle grid and sits on the brain’s surface (instead of being pushed inside), with each well containing engineered neurons grown from stem cells. The neurons are heavily modified, optimized for specific functions. Once the waffle-like device is installed, the neurons begin growing new connections – axons and dendrites – down into the brain tissue itself, forming biological links with existing neural circuits.

At least in the tests with mice, Science Corp. says it demonstrated that these additional neurons worked some of the time: Five of nine mice learned to move left or right when the device was activated.

“It does this in a perfectly bio-compatible way, because the brain is really just a bunch of neurons,” Hodak said. “Just neurons talking to neurons, the way evolution intended, save for the not-inconsiderable fact that some of the neurons come from a lab.”

What if something goes wrong? A patient can take a vitamin, “an FDA-approved thing that you wouldn’t otherwise take,” and the engineered neurons will die, Hodak says, describing it as a valve built into the biology itself.

What Hodak really wants

We’ve been talking for a while when Hodak reframes everything in one sentence. “I actually think BCI is a longevity-adjacent story.”

“The brain does two things: the brain is intelligent and it’s conscious. We know that intelligence is substrate-independent, because you get it in both brains and GPUs. But the end of the brain-computer interface quest, I think, is actually conscious machines.”

This is about cracking consciousness itself; understanding the physical laws that make subjective experience possible, and then engineering it into new substrates.

“In order to prove a theory of consciousness is right, you have to see it for yourself,” Hodak explains. “That will require these big brain-computer interfaces.”

Hodak thinks that once humans understand how billions of neurons bind together to create a unified experience — what neuroscientists call “the binding problem” — we can start doing truly wild things.

I almost hesitate to say some of those wild things include multiple brains working to form one consciousness. “You could really, in a very fundamental sense, talk about redrawing the border around a brain, possibly to include four hemispheres, or a device, or a whole group of people,” he says.

Hodak’s basically describing the plot of “Pluribus,” the new Apple TV show where an alien signal transforms humanity into a hive mind. It’s dystopian as hell. But Hodak seems to think the basic science is… sound.

“Will there be some giant super organisms that correspond to world cultures? Will there be dyads, like the next step up in marriage?” He’s genuinely uncertain how the technology will be used. “It’s kind of tough to imagine how it will get used, but I’m pretty confident those devices will get built.”

Basically, at the end of this path, you don’t have simply smarter humans; you have people who’ve merged with machines, with each other. Consciousness that spans multiple substrates, bodies, and minds.

“You could cure cancer, you could cure cardiovascular disease, you could cure all metabolic disease,” Hodak says. “But there’s this alternative view of substrate independence that just basically says, what if we didn’t need to solve those problems in the first place?”

What if, instead of endlessly patching failing bodies, we just moved the consciousness somewhere else?

The tipping point

What makes this conversation remarkable is how concrete everything sounds. Hodak isn’t hand-waving about “someday.” He’s got timelines, patient numbers, and regulatory pathways.

“By 2035, [biohybrid neural interfaces] will be basically available for patients in need,” he says. “And that will start to really deform the world in interesting ways.”

To be clear, Hodak isn’t saying healthy 40-year-olds will be lining up for brain surgery anytime soon. “These are [for] very serious brain surgeries,” he emphasizes. He does say that because people invariably age, “many people eventually become patients.”

Meanwhile, he claims the technology will improve, surgeries will get safer, and benefits will become more dramatic. And gradually, the patient population will expand. By the late 2040s — which is not so far away — Hodak thinks the tech will be “really ubiquitous.”

By 2035 is when things are expected to get weird. That’s when, Hodak predicts, “patient number one gets the choice of like, ‘You can die of pancreatic cancer, or you can be inserted into the matrix and then it will accelerate from there.’”

He tells a room full of people that in a decade, someone facing terminal illness might choose to have their consciousness uploaded and somehow preserved through BCI technology. The people in the room look both entertained and concerned.

Money, money, money

One thing that seems unlikely to change is that a small minority of people will have vast financial resources while the rest don’t. Currently, insurance covers treatments for macular degeneration patients. But as BCIs proliferate and improve, the entire economic model of healthcare may break down.

Hodak’s argument is that consumer tech has good deflation. Phones and computers get better and cheaper, so we buy more of them and spend more money; the market expands. But healthcare operates on what Hodak calls “a fixed bucket of money.”

The theory is that as BCI technology presumably improves and extends lives, there will be more things to spend healthcare dollars on. “The problem is that as new technologies come along that produce better outcomes and longer lives, there’s more stuff to spend money on for better outcomes,” Hodak says. “You can’t spend like 10 times as much on healthcare. This would be a catastrophe.”

“That is like a fundamental conflict that I think is going to eventually break the healthcare system as these technologies actually work in big ways.”

Honestly, at this point in the conversation, healthcare costs are far from top of my mind. But it’s an important point: either healthcare spending balloons to unsustainable levels, or whether someone gets a BCI is a matter of whether they can afford it. I can’t believe that class divisions based on cognitive enhancements seem like an actual, near-term economic reality, but imagine trying to compete against someone with perfect recall, or the ability to calculate something instantaneously. That would suck.

As our time is running out, I ask Hodak what he thinks might happen to society. He doesn’t have answers. “I worry a lot more about Twitter than I do about these things,” he says with a smile, adding that he’s more concerned about information manipulation through our eyes than through direct brain interfaces.

I leave the conversation thinking about “Pluribus,” which Hodak is yet to watch. In the show, the hive mind offers everything: perfect knowledge, freedom from loneliness, complete understanding. But once you’re in, you’re not you anymore.

I also think about how Altman’s comment, that he’d just ask the AI, tickled a room full of people because it sounded absurd to those outside that field.

Years ago, as a junior reporter in Silicon Valley, I had the same initial response to many ambitious ideas and products. A lot of those ideas have now taken over the world, so now I just listen and wait.