Disappointment takes two forms. The first allows you to preserve the fantasy of the thing you desired, which has merely slipped out of reach or maybe turned against you. The second reveals that the thing you believed in was never real in the first place.
Given the choice between two disappointments about the future of technology, the petulant children of the space age and their disaffected millennial issue are tempted by the consolations of the former. This disappointment says it is not that the jetpacks and teleportation they were promised were a fantasy, but that they will appear in their dark form, proffered by Silicon Valley overlords. Better to be betrayed than for the thing itself to have been illusory all along; even if we won’t be making it to Mars, at least someone will. From Black Mirror’s downloaded memories and digital slaves to Blade Runner’s biological traffic in bodies, even the most dystopic visions of successful technology are preferable to the melancholy of Jean-Luc Godard’s Alphaville (1965), a shoe-string budget sci-fi film shot in the future ruins of modernist Paris. In Alphaville, there are no dazzling futuristic technologies, not even evil ones. Instead, there are only the run-down remains of architecture built when people imagined the future would be different. That, plus a supercomputer watches your every move. The future looks like the present, only more so.
Our current situation is more Alphaville than Black Mirror: amid worsening ecological and infrastructural collapse, the only technology that seems to steadily upgrade are mobile devices offering escape into the virtual. Everything is falling apart by mistake except for your iPhone, which is falling apart by design. Given the lack of options, even a villain offering an alternative might look like a reprieve.
This is one way to account for the credulity that greeted claims Elon Musk made in his press event for Neuralink last month, even from those who consider themselves his critics. In the first public demonstration by the secretive neural implant company since a similar event held last July, Musk trotted out pigs who had for the past two months lived with a prototype of Neuralink’s device in their skulls. Flashing a slide that named diseases spanning addiction, memory loss, blindness, anxiety, and paralysis, Musk reasoned that each malady is caused by “electrical signals sent by neurons to your brain.” As a result, “if you correct these signals, you can solve everything from memory loss . . . to depression.” Later, in an aside too labored to be spontaneous, he said you could “think of [the device] as a Fitbit in your skull.”
Everything is falling apart by mistake except for your iPhone, which is falling apart by design.
In its current iteration, the “Link” is a chip the size of a silver dollar, implanted flush with the skull and attached to flexible electrode threads containing 1,024 channels that are “sewn” across the cortex, the outermost layer of the brain. The chip compresses information gathered by the electrodes about the brain, identifying patterns by listening in on the bursts of electrical activity known as “spikes” that occur when a neuron fires. Once the device has matched the in-vivo spike to its coded templates, it can pare down the “noise” of a cacophonous live brain into a digital “signal,” small enough to be transmitted by a low-bandwidth interface like Bluetooth.
This may sound impressive, but by the standards of current neuromodulation, the Link is underwhelming. Scientists have been recording the brain’s neural spikes since 1868, and using inlying electrodes hooked into computers since 1951. The cascade of Aphex Twin-style musical tones Musk played at the demo, representing “real-time signals” transmitted by the Link, can be created over a lunch break by anyone with a computer program capable of assigning a musical note to a numeric value. According to Andrew Jackson, a professor of neural interfaces at Newcastle University, 1,024 channels is hardly impressive, and the fact that the device could roughly predict a pig’s movement when walking echoes findings that have already been published. Even in terms of Neuralink’s basic near-term goals, a neuromodulation device called NeuroPace that went on the market in 2013 is already capable of sensing brain activity and intervening in neural circuitry in real time to prevent seizures.
But the hoopla was about science theater, not science. This splashy display was calibrated to lend plausibility to Musk’s long-term vision, a sci-fi fable that goes something like this: Neuralink will produce a consumer-facing neuromodulation device capable of regulating any psychological or neurological ailment—or, more to the point, augmenting the brain’s capability with computation in the face of an alleged imminent threat of superhuman Artificial Intelligence. It will be implanted in a fully automated procedure you can have done during your lunch hour, as simple as Lasik, that dispenses with costly surgeon’s and anesthesiologist’s fees, bringing it to an out-of-pocket price point. The system will be yanked out of the brain and upgraded every few years, like a phone.
In their haste to raise the obvious political and ethical questions such a technology provokes, most critics of the device have swallowed the premise that it could plausibly do what Musk says it will. If the above description sounds like “mind control” or “suicide of the mind,” that’s because it’s supposed to. Better for Musk to admit with faux coyness during a Q&A session that maybe his aspirations did, he guessed, sound “like a Black Mirror episode” than for his claims about how the device works to be dismissed as hype staged to drum up investment.
If the Neuralink system isn’t particularly innovative, neither is its imaginative horizon. The fact that the technology does not yet exist speaks to extreme technical and regulatory hurdles rather than the purported historical novelty of Musk’s vision; last month’s launch is far from the first time that elaborate claims about reshaping human nature through brain computer interfaces have been put forward since brain scientists began working with computers in the mid 1950s.
Elon Musk’s scuttled stage-business with cyborg pigs was, if anything, a more disappointing version of a public spectacle of science staged in bullfighting Cordova, Spain, in 1965 by Yale neurophysiologist José Delgado. That event was major news, carried by the New York Times in a cover story:
Afternoon sunlight poured over the high wooden barriers into the ring, as the brave bull bore down on the unarmed “matador”—a scientist who had never before faced a fighting bull. But the charging animal’s horns never reached the man behind the red cape. Moments before that could happen, Dr. Jose Delgado, the scientist, pressed a button on a small radio transmitter in his hand and the bull braked to a halt. Then he pressed another button on the transmitter, and the bull obediently turned to the right and trotted away.
The bull was obeying commands in his brain that were being called forth by electrical stimulation—by the radio signals—of certain regions in which the fine wires had been painlessly implanted the day before.
The Times rhapsodized that this was “probably the most spectacular demonstration ever performed of the deliberate modification of animal behavior through external control of the brain.”
Delgado was no fringe character, and he was skeptical of the lobotomists who were, at the time, accepted within mainstream brain research; he considered them to be crudely hacking at the brain’s magnificent architecture. Poised at the cutting edge of midcentury brain science, he was a rising star of what was then called neurophysiology. After being named the youngest recipient of Spain’s highest award for brain science, he trained with the renowned neurophysiologist John Fulton at Yale, and soon after secured a faculty position, going on to win numerous prestigious awards from institutions like the Guggenheim Foundation. Delgado had spent the 1950s building methods of stimulating and recording the living brain, moving from trying to find physiological correlates for movement to trying to map the neural levers for subjective states like hunger, sexual desire, pleasure, rage, and motivation.
The bull demonstration was his piece de resistance, showcasing his “stimoceiver,” a neural implant controlled through radio frequencies that could stimulate the portions of the brain where it was implanted. Over the following years, Delgado worked to adapt the stimoceiver for use in humans. By 1969, he had published his first results with human subjects, with funding for research provided by the U.S. Air Force, the Office of Naval Research, and the United States Public Health Service.
The device’s basic architecture was remarkably similar to the Neuralink implant, albeit with fewer electrode contact points. The stimoceiver functioned by coupling up to forty intracranial electrodes, planted throughout the brain, with a radio wave transmitter and receiver attached to the patient’s head. The signals from the depth electrodes were received by the amplifier on the patient’s head, which then controlled the frequency of the transmitter, relaying the signals wirelessly to the inputs of an EEG recorder and magnetic tape recorder located up to a hundred feet away from the patient. Simultaneously, the patients’ conversation and activities were recorded with sound equipment. The goal was to discover “correlations between electrical patterns [in the brain] and behavioral manifestations,” in order to give medical clinicians insight into the most effective neural target for psychosurgery in each patient. The neural patterns were analyzed by a computer to determine which neural region correlated with a given behavior.
As more newspapers and television specials covered his work, often using photos and footage disseminated by his lab’s press operation, the public began reading articles suggesting that Delgado was developing “mind control,” capable of radically transforming emotions and subjectivity. This was an impression he courted; in a comment typical of his style, Delgado remarked in a 1959 popular press article that “animals and humans may be controlled like robots, by pushing buttons.” With the publication of his 1969 book Physical Control of the Mind: Toward a Psychocivilized Society, Delgado laid out his case for embracing human engineering through reprogramming the brain. Evolution is constantly occurring, he argued; the next step must be for humanity to harness neural technology to reshape itself in the service of advancing the species, rather than annihilating itself through technological horrors like the atomic bomb.
If Delgado’s psychocivilized society rhymes with Musk’s aspiration to prevent human obsolescence by syncing the brain with artificial intelligence, it’s better understood as a political inversion than as a twin. As a staunch anti-Franco Spanish educated liberal, Delgado imagined human engineering as a natural outgrowth of social democratic governance, no different in kind than Le Corbusier’s public housing, or state vaccination campaigns. So it’s fitting that the anti-psychiatry movement of the 1970s, which took Delgado as its arch enemy, was laced with exactly the libertarian sensibilities that would, as historian Fred Turner has shown, eventually give rise to the politics of Silicon Valley.
But what the anti-psychiatry movement got wrong in their panic about Delgado is ironically the very thing that Musk is leveraging in order to claim that mind control has now arrived. Put simply, it’s much easier to manipulate motor functions in the brain than it is to influence (much less control) thoughts and feelings, and achieving technical mastery of the body’s movement does not imply that mind control is imminent.
Delgado’s spectacle with the bull relied on a key sleight of hand. While he claimed to have found the levers of motivation and subjectivity in the animal, the explanation was much simpler: his device was implanted in a motor area of the brain that made it impossible for the bull to continue walking forward. But the theater worked. The public was sold on the idea that mind control was on the horizon, while in reality the implant had only interfered with physical movement.
The neural correlates of motor control are easy to locate, found by poking around in the brain and zapping until the desired response is achieved. It’s much harder, if not impossible, to do the same with socially constructed feelings like “sadness” or “trauma.” So it’s not surprising that the first applications planned for Neuralink will be for movement disorders like paralysis. For a company fraught with internal chaos, described by former employee as riven with rapid turnover and wildly vacillating shifts in strategy, aiming at movement disorders offers a mercifully easy target, one that has the benefit of having already been achieved by multiple other research teams.
As for the possibility that Neuralink could make the leap from motor to psychological application, a thicket of technical and regulatory hurdles stands in the way. For one, the surgical robot it uses can only sew thin, flexible electrode filaments onto the surface of the brain, which largely controls physical sensation and motor function; it lacks the capability to implant them into deeper structures under the cortex. Since subjective and emotional experiences involve deep brain structures, it is likely not possible to influence emotion or mental illness from the cortex alone.
There is no universal biological substrate that corresponds to notions like “depression,” “sadness,” “anger,” or “joy.”
According to Alik Widge, a national leader in brain stimulation for psychiatric disorder, any attempt to use the surgical robot to implant deeper than the cortex will face two major engineering problems. First, you have to figure out how to get the delicate, floppy electrodes the Neuralink device uses into the deep brain. That aside, the robot relies on machine vision to dodge blood vessels, and a single mistake could be deadly. The idea that a brain implant surgery could be fully automated any time soon, Widge said, is a fantasy: “Implanting electrodes is easy as long as everything goes right, but you don’t want to see what it looks like when it goes wrong. So, could you do this in a strip mall with a guy who has a bachelor’s degree and a two-week training course? Maybe, once the technology is mature enough. But if you’re anywhere in the first ten thousand people to get one of these, you’ll want a brain surgeon standing by.”
David Darrow, a neurosurgeon who has worked for the past decade with neuromodulation systems, was more skeptical about the prospect of a fully automated deep brain surgery: “There’s absolutely no way. You’re never going to have anyone doing this but highly qualified surgeons, which creates an obvious bottleneck to making this a Lasik-style surgery.” If a single case goes awry, the device could be thrown into the same sort of regulatory limbo that halted gene therapy for a decade following an eighteen-year-old’s death in a clinical trial.
Beyond these formidable logistical roadblocks stands a more daunting epistemological problem. There is no universal biological substrate that corresponds to notions like “depression,” “sadness,” “anger,” or “joy.” These concepts are made out of language and politics; they are not universally shared between people or—more to the point—bodies. In Widge’s words, “Even if Neuralink were ready to implant tomorrow, we still couldn’t use it for psychiatric disease—because the science isn’t there to know what to do with it.” The problem is well-illustrated by the vicissitudes of research into using Deep Brain Stimulation, currently the gold standard of neuromodulation devices, to treat severe depression. “The word depression is almost useless” from a neurological perspective, Widge explained. “It doesn’t define a single biological entity.” The profound variation in what depression is is why some patients responded to the therapy while others experienced no change whatsoever. A large clinical trial seeking FDA approval for Deep Brain Stimulation for depression was shut down in 2013 when it did not show sufficiently promising results. In the wake of this failed trial, the major neuromodulation device makers have pulled out of mood disorder research and retrenched their investments in the safer realm of movement disorder.
Facing a Zeno’s pipeline from research to commercialization, the question is why Elon Musk would shoot wads of cash into an investment space defined by uncertainty. What is he getting out of putting up more than $100 million of the $158 million the company has raised?
Over the past decade, several startups sitting at the intersection of technology and a “real” service—think Uber or DoorDash—have received torrents of money from venture capitalists apparently unbothered by the fact that their investments have never turned a profit. Some commentators have hypothesized that Uber and the like are making a risky bet on automation, feeding their companies with infusions of investment capital while waiting for the competition, staffed by human workers, to die. But others have theorized that the business model is already working: the actual prize sought by capitalists is not immediate monetary profit but the data gleaned by platforms and devices. Some authors argue, then, that the data economy should be understood as a new permutation of capitalism in which data itself is not a commodity, but capital—that is, something capitalists want to own because it generates value through relations of exploitation.
A notable feature of the Neuralink prototype is that its electrodes are currently more suited to “listening” in on the brain, rather than stimulating neural activity. This makes sense given that the device’s most near-term use will be for paralysis, an application that relies on using computers to close the gap between the activity of sensory and motor areas in the brain and prosthetics. But it’s also what you’d expect from a company described by a former employee as attempting to be at once “a tech company or a medical device company.” As a tech company, the medical device arm of the venture doesn’t need to turn a profit or even be very widely used in order to yield data about neural activity that will legally belong to Neuralink.
Real-time information about neural activity is currently one of the hardest forms of data to acquire: everyone has a phone, but very few people have neural implants. This is why patients with Deep Brain Stimulation implants are treated as precious resources by scientific researchers; they are often work simultaneously multiple research teams running experiments in which the brain data gleaned by the device can be coupled with behavioral data. By combining different forms of data—the sort of information your phone collects, for example, and cortex activity—both sets become more meaningful. That is, they become more useful in predicting and directing behavior.
The question is why Elon Musk would shoot wads of cash into an investment space defined by uncertainty.
While it’s not immediately obvious what business models will emerge to capitalize on neural data, a rough shape of the answer is suggested by Rune Labs, a tech venture founded by an alumnus of Alphabet’s bioscientific wing Verily Life Sciences. Most medical device manufacturers are behemoths of the old economy, lacking the resources to curate the vast amounts of data their devices generate. Ditto for university researchers, who rarely have the margins in their research grants to purchase or build the computational tools necessary to correlate large quantities of behavioral and neural data. Enter Rune Labs, which offers device manufacturers and researchers a deal: give us access to the data generated by your neural implants, and in exchange, we will provide access to state of the art data storage and computation.
From their end, Rune Labs has developed a variety of phone apps to glean data about self-reported mood. (Similar research is ongoing in “digital psychiatry” to build apps that harvest data about everything from voice modulation to exercise which can then be coupled to the information about brain activity gleaned from neural devices.) The only restriction on Rune’s use of neural device data is that they have to keep patients anonymous. More and more, this looks like the business model that will define neural implants. As Alik Widge remarked, “The idea that your data is the product is already here with brain implants. Neuropace has already said that they’re moving toward being less of an implant company and more a brain data company.”
Of all the wild speculations Elon Musk made during the Neuralink launch, the most accurate prediction was his quip that the device is “sort of like if your phone went in your brain.” “Sort of like,” indeed: Neuralink is like a phone in that it is yet another machine built for generating data. While the device does not represent a major advance in brain-machine interfaces, and the pipeline for applications beyond movement disorders is at best decades long, what Neuralink does offer is an opportunity to harvest data about the brain and couple it to the kinds of data about our choices and behaviors that are already being collected all the time. The device is best understood not as a rupture with the past, but as an intensification of the forms of surveillance and data accumulation that have come to define our everyday lives.
Technology critique should not buy into the hyped fantasies people like Elon Musk peddle; mind control is no closer within reach than any of the dreams of the futuristic past, like flying cars, or socialism. What we have to worry about is something much more disappointing—that the future will look like the present, only more so. In that way, Musk is right: the future is already here.