What is Neuralink and why does Elon Musk want to get inside your head?


Staff member
Neuralink is a company that is developing brain-computer interface (BCI) technology. Founded by Elon Musk, the intermediate goal is to provide medical technologies that will help people who suffer from brain disorders. The end goal is to provide a method by which humans can beneficially co-exist with artificial super intelligence.

Wait, what? Back up. Artificial super intelligence? Like, talk to the cloud with our minds?

When Elon Musk famously described artificial super intelligence (ASI) as “summoning the demon” he was alluding to his concern that ASI may not turn out to be helpful or benign. If it’s only benign, the best case scenario is that an ASI might regard humanity as a creature at the ‘pet’ level of relationship. We would be no more significant to the ASI than a pet cat is to us. If it becomes malevolent? We’ve all seen The Terminator.

Tim Urban wrote an exhaustive article about Neuralink at his website Wait But Why. That article starts in a timeline near the beginning of evolution, then takes a slow walk through time to get to where we are now. One incisive observation is that unless we are at the top of the evolutionary tree, then we are just like everybody else below — where things aren't so great — because we are on top. Do we want to risk losing that status to an ASI?

It’s a fascinating read, but I was most interested in the business methodology that could bootstrap humanity into a symbiotic relationship with ASI, and the possible implications of how that would actually work. Tim Urban explains the business model that Elon Musk uses to bootstrap his companies from an initial business model to one that reaches for a greater goal.

Here it is in generic graphic form:


Here it is for Neuralink:


(Images are of my own making, but are adapted from original graphics at waitbutwhy.com)

The intermediate goal is to get to that purple box — light the match that ignites an industry. That requires innovation at the company level, which is where Neuralink comes in.

The first major challenge is bandwidth — getting data out of and into the brain at a useful rate. There are 20 billion firing neurons inside the cortex of the human brain. Current technology is able to record the electric state of 500 neurons. About 1 million neurons must be monitored before a useful BCI can be implemented. If the rate of increase in the number of monitored neurons follows a Moore’s law pattern and doubles every 18 months, it will take until the year 2034 to reach 1 million.

Looking at various points in time over the last 50 years, Ian Stevenson and Konrad Kording published a paper that documented the maximum number of neurons that could be simultaneously recorded at each point on that timeline so they could develop a trendline. That research indicates that the historical rate of doubling for BCI bandwidth was every 7.4 years. Starting at 500 neurons, at that rate it would take until the end of the century to get to 1 million. To monitor every neuron in the brain would not happen until the year 2225.

The needed technology breakthrough (the purple box in the diagrams) is to get on a path that is closer to a doubling every 18 months rather than 7.4 years.

The second major challenge is implantation. Current technologies require surgery, and some end up with a wire coming out of the skin. That is not a technology that a lot of people will embrace. Neuralink plans to work on devices that will be wireless, which in itself brings a lot of new challenges. To send and receive a lot of data wirelessly, the implant has to manage signal amplification, analog-to-digital conversion, and data compression. The device will also have to be bio-compatible so it can withstand the environment inside the skull and not trigger the body’s immune system. How such a system would be powered is also an open question.

Huge challenges, also contained in the purple box in the diagrams.

But let’s say Neuralink is successful and the society at large embraces the idea of connecting our minds to the cloud — that we each become a node on the internet of things. How would that actually work? Tim Urban interviewed several leaders in the industry and collated their responses to posit four possible scenarios. Each level is an incremental step up the digital experience ladder.

Level 1: I want to know a fact. I call on the cloud for that info — like Googling something with my brain — and the answer, in text, appears in my mind’s eye. Basically what I do now except it all happens in my head.

Level 2: I want to know a fact. I call on the cloud for that info, and then a second later I just know it. No reading was involved — it was more like the way I’d recall something from memory.

Level 3: I just know the fact I want to know the second I want it. I don’t even know if it came from the cloud or if it was stored in my brain. I can essentially treat the whole cloud like my brain. I don’t know all the info — my brain could never fit it all — but any time I want to know something it downloads into my consciousness so seamlessly and quickly, it’s as if it were there all along.

Level 4: Beyond just knowing facts, I can deeply understand anything I want to, in a complex way. ... Could I download Moby Dick from the cloud into my memory and then suddenly have it be the same as if I had read the whole book? Where I’d have thoughts and opinions and I could cite passages and have discussions about the themes?
- Tim Urban, WaitButWhy.com

Any one of those would be great, but in addition to ‘knowing facts’ there are additional levels of experience that could also be recorded, replayed, or directly experienced. What about emotions? The senses? Thoughts?

Imagine a game where the players ride actual flying jetbikes through a physical reality. They are fifty stories high above the ground, and riding through the airspace between futuristic buildings. But that’s not all they (and you) see. Since they (and you) have a BCI system installed, you are both able to experience the virtual reality that the game engine maps on top of the physical course. It might be a jungle setting, or an ocean reef, or a grand canyon. As the players ride their jetbikes, you can feel the wind in your face, the thrum of the jetbike beneath you, and the emotional thrills and chills of the competition. You can ‘hear’ the players as they discuss strategy via technology-assisted psychic connection. You can feel it when they get hit by virtual obstacles. And when it’s all over, the experience is part of your permanent memory as if you did it all yourself.

Don’t feel like playing a game? Maybe you can ‘drop in’ to the mind of actual tourists and virtually travel in their mind with them as they visit exotic places. Don’t have time to do that in real time? Then you can buy the recording of the vacation and install it into your own memory. Ahhhh… That was a great vacation!

As we all know, with great power comes great responsibility. Such a system is tailor-made for abuse. If memories can be implanted, how can you be certain that your memories are your own? How do you know if your prejudices didn’t come from somewhere else? Why do you always want to buy that next new thing? Aside from the usual suspects of oppressive governments, giant corporations, and terrorist hackers, how could we trust the artificial super intelligence that would run such a system to not use it against us?

Every new technology brings with it the risk of certain abuse. In every case, most people arrive at a point where they accept the risks in order to enjoy the benefits. Very few want to turn the clock back even 20 years, much less 50 or 100 years. Most of us agree that the benefits of modern technology outweigh the risks. Given the existential risk that comes with artificial super intelligence, we are likely to accept the risks that come with brain-computer interfaces.

Our only other option is to become akin to a pet cat.

Ultra Mod - Chapter One

Click/tap image to read the first chapter of Ultra Mod.