By Nima Bidari
Elon Musk’s path to become real-life Tony Stark continues with Neuralink. Founded in 2016, Musk planned to take human perception to the next level with brain implants that can read and write action potentials (i.e. brain signals) like computer code. Although the development of the technology has been incredibly secretive so far, we know a few things: They do not claim to be and are not the first to consider this, and scientists have been and will continue to work on this technology for decades. In fact, we’ve been doing it since the ’50s, reading the brain signals of rodents and large mammals, and even allowing apes to control prosthetic limbs with them.
Moreover, Neuralink wants to merge human and artificial intelligence. What this means exactly, is not clear yet, but fantasizing about the possibilities quickly leads us into a rabbit hole too exciting to get out of. Imagine playing symphonies and movies in your head, manipulating emotions and fears, and waking up with a brand new set of skills, like playing the piano. Imagine if we could have Alexa or Siri installed into our brains, or download Wikipedia onto our brains. Of course, there are limitations, and it is no longer a factor of how or whether we can do it, but what even is “it”?
Scientists still do not understand a lot of how memories are formed and stored, and we especially have no clue how to manipulate fears and emotions, safely. Consequently, we have to imagine a less exciting, but much better world where blindness, paralysis, deafness, and mental illness are a thing of the past. So, where does Neuralink take its place in the world of neurotechnology?
Apparently not at the front, or even close to it. Electroencephalography (EEG) has been commonplace for years now, tracking brain wave patterns via electrodes on the scalp, meaning it can detect seizures and night terrors, and detect brain wave abnormalities and caused by tumors. The problem with EEG hardware is that despite their safety benefits, the signals are noisy meaning there is no way to read complex thoughts and memories. To overcome this, scientists have used brain implants, connecting microelectrodes directly to the brain. To give you an idea of how old this technology is, Ida Hyde invented microelectrodes in 1921, and this type of brain-computer interface has existed for decades without Musk’s ambition. Although Neuralink has 1024 electrodes, in contrast to BrainGates 100, they still have little to show for their massive 160 million dollar budget, 100 million of which was provided by Musk himself. Their August 28th “product update” did not show the “Matrix in the matrix,” Musk promised on twitter. Even Musk downplayed it, referring to it as, “kind of like a Fitbit in your skull.” Nevertheless, this is Elon Musk we are talking about. Nothing he has ever done was like the Model T was to wooden carriages, and instead he took something that existed and made it way better.
If you missed the product update, here is a small recap: The chip is implanted on the top of the head, flush with the skull, and is charged wirelessly, with induction coil technology. Furthermore, Musk seems to be keeping up with his promise that you will be able to walk into a doctor’s office and a robot will install it within an hour. He unveiled a futuristic and intimidating looking robot meant to use an electric saw to cut a hole in your skull, and place the electrodes and device. According to Musk, the device can be removed for any reason, including upgrades, as well. In terms of functionality, the device is unimpressive. It can detect and predict the pig’s limb location and detect if the pig hits its snout against a wall.
Many critics say that Musk has no idea what he is doing, and is just throwing money at it. They do not seem to realize the clear trend with Muskian technology. First, he spends a lot of his own money to keep the company afloat, trying to stay private for as long as possible and usually overestimates his developmental and manufacturing speeds. He is optimistic, to say the least. Despite missing most of his deadlines, he never over promises in terms of functionality. A 500km+ range for an electric vehicle was unheard of, and charging it fully within half an hour is beyond what anyone thought was possible, with lithium-ion batteries. Even though the Roadster was no more than a life-size Lotus Elise RC car, he turned Tesla into autonomous car pioneers and spearheads the electrical vehicle industry.
So clearly Musk overpromising on deadlines is nothing new, and we should not downplay what Neuralink has done so far, simply because they do not seem to be on track for human trials in 2020. As Musk said, “The idea of going to the moon is trivial, but going to the moon is hard.” However, there are other reasons to doubt him. We are not even sure specifically what thoughts and memories are, let alone if it is possible to read and write thoughts.
Most brain-computer interfaces today use a technique called biomimetic decoding to translate action potentials to computer code. Essentially, brain activity is recorded while the user imagines using the left or right arm, and the computer learns which neurons most often predict each arm’s movement. Then, the prosthetic is coded to move when those neurons fire. This has even been used to try and drive a car by imagining the car move. As with all things, there are limitations including both the hardware and the fundamental concept. 1024 electrodes are sufficient for prosthetics and if Elon delivers on his promise of increasing that tenfold, then treating blindness and deafness is not out of reach for Neuralink’s capabilities, but complex thoughts and memories often involve billions of neurons in complex circuits, thus the learning period for the computer would be too long, and require more computational power.
More importantly, scientists say that our current understanding of the brain makes uploading and downloading thoughts impossible. And that’s only in reference to the parts that we do understand. Moreover, microelectrodes often stimulate more than one neuron at a time, so in basic terms, we can stimulate a flash of light easily, but we can not create a high-resolution image in our brains.
Despite all this, we should remain optimistic, if a little reserved about Musk’s outrageous deadlines. Andrew Jackson, a Newcastle University professor of neural interfaces, is pessimistic about Musk’s deadlines but believes that Musk is not necessarily overpromising. He said in an article for The Conversation, “That said, decades of research have shown that the brain does not yield its secrets easily and is likely to resist our attempts at mind hacking for some decades yet.” In addition, many other professionals in the field downplay what Musk has done so far, but Musk responded, “It is unfortunately common for many in academia to overweight the value of ideas and underweight bringing them to fruition. ” Intriguingly, brain-computer interfaces began development right after the moon landing. It seems that we may be decades away from telepathic communication and virtual reality beyond Ready Player One, but one thing is clear: Humanity has found the next “giant leap for mankind”.