We were talking about the stifling amount of transhumanist-themed sci-fi shows on Netflix et al. when a friend of mine, who is a big fan of Elon Musk, brought up Neuralink - perhaps these fictional worlds are not as distant as we think.
Not knowing anything about Neuralink aside from a couple of soundbites on the radio, I decided to read up about it a little more, and try and understand how Elon Musk was going to create a “Brain-Machine Interface”.
The Neuralink White Paper authored by Elon Musk and Neuralink (?) is pretty accessible, and a worthwhile read for anyone wanting the nitty-gritty regarding the technical implementation of one example of a Brain-Machine Interface (BMI).
In general terms, Neuralink’s offering, although currently in development, represents multiple dense arrays of electrodes fixed on polyimide polymer “threads” that are then surgically inserted into the brain. The electrodes pick up action potentials from the neurons that surround them, relay the information to a connected chip for amplification, digitization, and processing, then transmit it wirelessly to a Bluetooth device worn on the ear.
The purpose of the presentation in July was to recruit more talent to bring Neuralink from its current prototype phase to what it’s eventually meant to be - an interface between brains and machines that would, in Musk’s words “achieve a symbiosis with artificial intelligence”.
What seems very promising in the short term, is the ability to understand and potentially cure neurological disorders. Neuralink aims to be able to apply algorithms to electrocorticograms produced by electrodes inserted into the motor cortex in order to decode them into a comprehensible signal reflecting what physical movement that brain is trying to make. With further development, that will mean that despite having a damaged nervous system incapable of relaying these signals to the target muscle groups, Neuralink will allow the brain to potentially control prosthetics to recreate the intended movement. Eventually, Neuralink could integrate with muscle stimulation technology to give users back control of muscles that have long been unavailable to them.
It’s hard to say what the disadvantages of this kind of medical breakthrough could be. It really seems as though this could significantly improve the quality of life for people affected by a whole range of disabilities, and allow them to contribute meaningfully to society. Most likely it will first be made available primarily to those willing to pay for it, although Musk thinks that given the strain placed on the healthcare system by individuals with disabilities, subsidizing Neuralink makes perfect economic sense.
What’s really interesting is what happens when this idea is developed along its logical progression. The presenters from Neuralink mention that reading data out of the brain is just the first step - it should also be possible to put information into the brain. This is already being done for cochlear implants, but it could also take the form of artificial vision, made possible by stimulating neurons in the visual cortex in just the right way.
These would be applications that restore human faculties to those that have been deprived of them for one reason or another, but what about enhancing humans with faculties they’ve never had? What about a slow, methodical, incremental enhancement of Neuralink that does away with smartphones in favour of a more integrated and capable human being? What about having your ordinary experience of reality integrated with overlays containing any kind of information imaginable? You can imagine having satellite feeds piped directly into your brain, weather data, traffic data, social media feeds, stock reports, location feeds, whatever you want - customized how you want and available for your consumption. An enhanced memory would make your experience of the world not just what you see in front of you now, and not just the sporadic highlight reel we’re accustomed to, but a rich, high-fidelity recording of everything we’ve ever seen, available for replay, stored externally but projected directly into our minds. Additional neurons for enhanced processing power would allow us to construct concepts at a much more granular level than those captured by the coarse weave of our current wetware substrate. This would be what a solution to Musk’s bandwidth problem looks like - every one of us, a Great Human.
One compelling (but ultimately unsatisfactory) perspective on what it would mean to be an enhanced human being is afforded by the example of Gottfried Wilhelm Leibniz - the last person to “know everything”. Leibniz made major contributions in several branches of mathematics, inventing calculus independently of Newton, as well as physics, philosophy, and pretty much all other major natural and social sciences.
In addition to being incredibly gifted, it’s worth pointing out that he had ample material to help his intellectual development. His father was a philosophy professor at the University of Leipzig in what is modern-day Germany, and Leibniz had free access to it since he was 7. Most people at that time didn’t have libraries filled with advanced texts on philosophy and theology, so it’s interesting to think about what Leibniz’s achievements would have amounted to if, all things being equal, he didn’t have a library’s worth of books to sink his teeth into during his most formative period. Suffice it to say that in 2019, we’ve already discovered the importance of data. With the final Neuralink, the amount of data available for immediate consumption to any individual would be unfathomable.
Leibniz believed in the possibility of an “alphabet of human thought” - a system where every concept humans can come up with could be assembled from a set of primitives. In this system, the grammar governing propositions constructed using the alphabet would mirror logic itself, resulting in the possibility that the validity of an ethical or metaphysical claim could be determined by whether or not it could be spelled out using the alphabet.
It is obvious that if we could find characters or signs suited for expressing all our thoughts as clearly and as exactly as arithmetic expresses numbers or geometry expresses lines, we could do in all matters insofar as they are subject to reasoning all that we can do in arithmetic and geometry. For all investigations which depend on reasoning would be carried out by transposing these characters and by a species of calculus. (Preface to the General Science, 1677)
I’ve always been fascinated by the idea of looking at the world and coming up with a concept that is not expressible in any existing system. Perceiving the world means carving it up into concepts, which we can share with each other using language, or other kinds of symbol systems. Coming up with new systems for doing this happens all the time, as in the formation of Creoles. And yet, Leibniz, forgetting for a moment the aforementioned discovery of calculus (and coming up with a notation for it), was trying to do this for any thought a human being could have.
With a “final” Neuralink, every one of us would be a Leibniz, perceiving patterns in reality that would be inexpressible in our current languages, necessitating the invention of new, much more intricate and subtle symbol systems - we would become incomprehensible to modern humans.
In Amusing Ourselves to Death, Neil Postman talks about how television is trivializing communication due to a dilution of the ratio of information to action. He argues that this process really began with the telegraph, when it became possible to cheaply and reliably transmit information across vast distances, stripping it of its original context, thus:
Telegraphy gave a form of legitimacy to the idea of context-free information; that is, to the idea that the value of information need not be tied to any function it might serve in social and political decision-making and action, but may attach merely to its novelty, interest, and curiosity… In both oral and typographic cultures, information derives its importance from the possibilities of action…but the situation created by telegraphy, and then exacerbated by later technologies, made the relationship between information and action both abstract and remote.
If a version of Neuralink, in its ultimate form, were available, Postman’s argument no longer applies. We would have infinite context - each additional piece of information at our disposal contextualizing every other. And with additional information would come additional inputs for possible decisions along with knowledge of how to realize the outcomes we would want to see in the world, until the information action ratio becomes 1.
The real question that emerges from all this is who would get the treatment that would allow them to use Neuralink? At some point in the future there would be a critical mass of users who had undertaken it, and those who hadn’t How would we decide who goes in which category? Or would it be mandatory from birth?
If enhanced human intelligence produces new, previously inconceivable ways to improve itself, resulting in a still greater intelligence which becomes the next input for the same process, would it really be possible for ordinary humans to coexist with whatever would result from Neuralink?
Elon Musk claims that this will just be optional - but how realistic is this? This isn’t exactly the kind of quaint fascination we feel when we meet someone who still uses a fax machine for personal communication.
This article from The Conversation mentions the possibility that:
Cognitive enhancement will join coffee, painkillers, antibiotics and even smart phones in becoming the commonsensical and expected choice. The burden of proof will have shifted such that everyone will be expected to enhance themselves; justifications will have to be given to explain why we should not comply.
Judging by all that speculative fiction (on screen and in print) I mentioned in the beginning, I get the general sense that people are more concerned about the problems than excited about the possibilities. I guess that makes sense, humans are loss averse, and while we don’t really have any concept of what we stand to gain from a Neuralink 2.0, we have a very real sense of what we’re putting on the line.
Despite working in tech, I’m not very optimistic about our prospects. I’m certainly not an AI researcher, so from one naive point of view, teaching human values to a machine seems like a colossal undertaking - the acute futility of translating a joke, while the wellbeing of your species turns on a smile.
Douglas Rushkoff, creator of the excellent Team Human podcast encourages us to confront the idea that for us, our future has become “less a thing we create through our present-day choices or hopes for humankind than a predestined scenario we bet on with our venture capital but arrive at passively”. Which is to say that Neuralink, or whatever other guise the Singularity takes, is not at all a consequence of choices we make now, but an inevitability that we are condemned to encounter, regardless of how circumspect we are in our groping attempts at innovation. It may as well already have happened.
His thesis is that we should not be looking at the world in this way. We know we don’t have to reach far for examples of tech that started out with the best intentions only to become cobblestones on the proverbial road. But we should be taking responsibility for our creations, and thinking critically about how they facilitate meaningful, fulfilling interaction, rather than the next IPO.
Despite a healthy skepticism, I absolutely love this message. I want to be a part of the Team Human he’s talking about, but it’s difficult to imagine what that looks like in today’s tech landscape, let alone a future one. I hope that when the time comes, before we’re whisked away by the emergence of a new species, Neuralink will grant us the imagination we need.
“Gottfried Wilhelm Leibniz.” Stanford Encyclopedia of Philosophy, Stanford Encyclopedia of Philosophy, 22 Dec. 2007, https://plato.stanford.edu/entries/leibniz/.
Jolley, Nicholas. The Cambridge Companion to Leibniz. Cambridge University Press, 1995.
Musk, Elon, and Neuralink. “An Integrated Brain-Machine Interface Platform with Thousands of Channels.” BioRxiv, 16 July 2019, doi:https://doi.org/10.1101/703801.
Postman, Neil. Amusing Ourselves to Death: Public Discourse in the Age of Show Business. Penguin Book, 2005.
Rushkoff, Douglas. Survival Of The Richest. Medium, 5 July 2018, https://onezero.medium.com/survival-of-the-richest-9ef6cddd0cc1. (Premium Medium article)
“The World’s Most Valuable Resource Is No Longer Oil, but Data.” The Economist, The Economist, 6 May 2017, https://www.economist.com/leaders/2017/05/06/the-worlds-most-valuable-resource-is-no-longer-oil-but-data.
Vincent, Nicole A, and Emma A Jane. “Put down the Smart Drugs – Cognitive Enhancement Is Ethically Risky Business.” The Conversation, The Conversation, 15 June 2014, https://theconversation.com/put-down-the-smart-drugs-cognitive-enhancement-is-ethically-risky-business-27463.