What exactly can Neuralink do? While SpaceX regularly sends rockets to orbit and Tesla stock has its own way of defying gravity, the weight of Neuralink will be felt by all, through space and time.
Neuralink is Elon Musk’s Neurotechnology company which develops brain-machine interfaces. These devices have use cases ranging from treating Parkinson’s Disease and restoring bodily function in paralyzed patients, to control computers with your thoughts and communicating telepathically with other users.
All of this is towards the ultimate goal of facilitating a symbiotic relationship with artificial intelligence. And while the benefits of AI becoming Cortana from Halo seem great, the risk of it becoming Skynet from The Terminator is extremely asymmetric.
Neuralink’s ultimate goal is to form a digital brain “layer” that works in cooperation with existing layers in our natural brain structure. The term “layer” refers to an organization of the evolutionary timeline of the human brain. To fully understand this, one must understand a bit about evolution and neuroscience.
When the first amphibians arose and began to walk on land, they began to develop what we can recognize as the first brains. These brains evolved to become the brainstem in humans. These primitive amphibian brains, and now the brain stem, control mostly unconscious life-sustaining functions like breathing and heartbeat regulation.
We will call this layer the survival layer. Soon, mammals developed limbic systems, which we have to this day, involved in more complex tasks like memory and emotion. We can refer to the limbic system as the animal layer.
Millions of years later, when the last primates evolved into early hominids, a new “brain” began to form known as the neocortex, the human layer.
So how does Neuralink connect to our current brain layers? Neuralink bypasses the long timeline of evolutionary biology in which brain layers take millions of years to develop and gives us a fourth layer – a cyborg layer. A surgical robot implants a “Link” inside your skull and threads thousands of wires that can sense when neurons are communicating with each other through electrochemical signals.
The Link picks up these signals from neurons and utilizes machine learning to decode and encode new signals that will communicate with other neurons, directly influencing brain function by affecting the messages these neurons send and receive.
With a programmable cyborg layer, the applications are almost endless. The most immediate use case is in the medical field. Paralyzed patients can use their Link to wirelessly send signals to their limbs to bypass their injured spine. Patients with permanent blindness can use artificial retinas that send light information directly to the Link.
The Link can process this information and encode it into neuronal signals that mimic those sent by natural retinas. Mental illnesses like ADHD and OCD can also be treated as the Link can effectively measure the number of neurotransmitters in the brain and adjust their production as necessary.
There is more to Neuralink than just its healthcare angle. Neuralink can not only fix what’s broken, but it can give us more knowledge and computational ability than we could have ever achieved with just our prefrontal cortex human layer.
With access to the internet, the Link can begin to transmit, store, and process information at inconceivable speeds. Imagine a world where you never forget anything because all of your thoughts are backed up to the cloud. Instead of phoning your friend, you can just beam your thoughts to them. You can now describe to someone the exact taste and texture of Parisian foie gras without a single word.
Consider a new trading strategy; you can now visualize the backtest of the portfolio and adjust the model by just thinking it. Just as Alan Turing could never predict the iPhone, we likely cannot even conceive of the outlandish capabilities of Neuralink.
While it is easy to conceive of the importance of Elon Musk’s other companies – Tesla and SpaceX – in their fight against climate change, Neuralink’s mission is more nuanced. At first glance, Neuralink aims to cure some of the debilitating diseases that plague humanity. However, Neuralink is fighting another, more sinister threat that may completely blindside us; artificial general intelligence.
Let us go back to our evolution discussion for a moment. Humans have become the apex predator on Earth because of our neocortex which allows us to think. Other animals are bigger and faster but humans’ unique ability to reason and work together has given us an unfair advantage. If AI can think, then why wouldn’t they replace us just as we have every other species?
Sure, AI in its current iteration is not HAL 9000. But that doesn’t mean we should take it lightly.
AI has already replaced millions of jobs in the manufacturing sector. Low wage service jobs like fast-food workers and bank tellers are likely next on the chopping block as natural language processing becomes more commonplace. It is only a matter of time until knowledge workers like investors and software developers are replaced by bots that don’t need to be paid and don’t eat, sleep or quit.
And these changes are only accelerating.
Unlike climate change which we are trying to avoid, we are actively investing billions of dollars and millions of labor hours into developing artificial general intelligence systems. And progress is being made exponentially.
In the 1980s, Gary Kasparov, then the best chess player in the world infamously declared that an AI will never be able to beat him.
A few years later, IBM’s first AI bot lost to Kasparov twice, seemingly confirming his theory. In 1996, Kasparov played IBM’s new program, Deep Blue, and won 4 games out of 6; he had been proven wrong.
Just the next year in 1997. Kasparov played the same program and lost 3 games and had 1 stalemate out of 6; the AI was competitive with the best.
Fast forward more than 20 years. In addition, the reigning champ Magnus Carlsen will get swept 100-0 by current deep neural network implementations. Even at games like GO, which are objectively harder given the search space of moves to make, AI can beat grandmasters.
At the extreme cutting edge, Google’s AlphaStar model can beat 99.8% of human players at StarCraft II, a real-time strategy game. If AI can progress this fast at games, how long until it can beat humans in every task?
In conclusion, this poses a series of existential issues. How does an economy function when almost all jobs are done better by robots that continually improve themselves? How do we compete with a “species” that can evolve in a matter of minutes compared to our million-year long feedback loop? And what do we do if they decide that we are in the way?
If you cannot beat them, join them.
Our only choice is to fuse with AI.
Neuralink’s thesis is that by creating a pathway with enough bandwidth to communicate with algorithms, we can hope to form a symbiotic relationship with AI. Lastly, there we can leverage the computer revolution. Furthermore, our general intelligence to ascend to the next stage of our evolution. Or at least, not become a house pet to our AI overlords.