Think of the computer sitting on your desk in front of you. You fire it up, you launch your email software, you use its interface to create an email to your boss saying you’re at home sick. Well, you’re not at home sick, in fact, but sipping your bubble tea in a nearby cafe, busy catching up on your favourite tech blogs. Doesn’t matter. Yet think about the relationship you have with the machine at this moment. At every moment of sending that email you are perfectly aware who you are, and by looking at the computer interface you can tell whether the machine is receiving your cues when you type the text in. You can tell that you manifest your behaviour by clicking the “send” button on that email client, and that’s where your action ends, and machine action begins. In a way, there is a certain established pipeline of intention coming from your brain, to your physical movements that gets translated by the interface into the machine commands. These commands then convert into electrical signals, and so on, all the way to the electric signals generated by synapses in the brain of your boss. She’s inevitably getting upset to hear about you being sick and perhaps a little annoyed about the project deadline being missed today.
Now, welcome to the era of Neuralink. Founded in 2016, Neuralink is Musk's neural technology company tasked with conjuring up an implant designed to interface directly with the human brain. The initial goal of the implant, says Musk, is to use it as a means to treat brain injury and trauma (Business Insider). With one of those Neuralinks installed, you wouldn’t have to use your body parts to type anymore. Instead, the machine that has been surgically injected into your brain reads all your electrical signals directly. It then sends that email without you ever touching the keyboard - provided, of course, you can pronounce the words in your mind without spelling mistakes. However you look at such a human-machine interface, it comes out pretty scary, huh? Robots tinkering on the inside of your skull, you know… But let’s look more closely at the main criticisms of Neuralink to see if it’s really as bad as we’d usually imagine.
1. Psychological and physiological implications
In a well-weighted ethical critique of brain-machine interfaces (BMIs), Director of the Arizona State University Risk Innovation Lab Andrew Maynard points out a few important things to keep in mind. There are plenty of things that can go wrong when you use laser beams to melt a “neural lace” - a term, in fact, adopted from Iain M. Banks 1996 sci-fi piece, Excession, see the book excerpt in the blockquote - through the skull and into the brain. Yet despite the risks, for patients and their families these will be most likely outweighed by the tremendous benefit of being able to become a functional member of society again. Disabled people will be able to use their paralysed limbs, be able to see, hear - or perform any other essential functions.
2. Obsolesence
Now, that’s where the real trouble begins. A big real-world complication that Neuralink is facing even now, before its launch, is obsolescence. In 2019, Musk claimed that this year we will see first human patients - while in Joe Rogan podcast released this month, the deadline has moved to 2021. Technology moves rapidly, particularly around BMI, and seemingly banal questions like “how do we upgrade the hardware when the new version of software comes out?” can be a show stopper. New and more effective versions of software that we develop are going to require new hardware. Yet if we need to perform brain surgery on millions of people, who are inevitably ageing and whose health was not so good to begin with, it makes shipping of new software releases, reasonably, quite problematic. In other words, undergoing brain surgery is slightly trickier than getting a new iPhone. This could be one of the reasons why Musk is hesitant to implant the hardware in its current shape. Essentially, while the humankind struggles to come up with a USB port that is going to be one and for everybody and for all time, similar to the human ear or mouth, no effective BMI can be implemented as a long-term scenario.
Speaking of software development, it should be noted that Neuralink is currently hiring software engineers, big time. If you’re full stack or interested in machine learning, you should probably apply. There has been a discussion on Reddit as to what are the potential programming languages, technology stacks, architectures that will be used for Neuralink apps. It is, it seems, essentially Webservices peppered with a great deal of Rails, React, Typescript and Postgres. At the bottom of the stack it most probably relies on languages with efficient signal-processing and machine learning capabilities, such as Python - see another related Reddit. The bottom line, whichever way the brain activity is translated into a control signal, it will be up to more conventional software implementation to process it.
3. There will be two classes of humans, one with AI-enhanced brains and the lower class
This is not likely to happen for two reasons. Firstly, we live in the knowledge economy, which makes profits on users’ data, and the aim of any new or existing business venture is mainly to collect the user data. Once the data is collected, the technological tools are then used to analyse and create business strategies based on that data. Quite logically, these new strategies are aimed at capturing even more data in the easier way. With this idea in mind, it wouldn’t surprise us if Neural Lace, once released, is going to hit the market running and will be offered cheaply to everybody. Even for those who completely struggle to buy it, there will be a monthly instalment plan and all sort of deals. After all, it’s just another tool of collecting user data, and it needs to be out there in order to generate any revenue.
The other reason why the new technologically-privileged class is not likely to appear is vagueness around the device’s security and ethics. What kind of privilege is that anyways, to be exposed to neural emotionally-enhanced ads that effectively turn you into a zombie, and use pretty much all of your brain’s contents for marketing and training purposes? Let alone forgetting to update the software and risk your brain being hacked into.
To sum up
With $100 million of investment from Musk himself and about a half less from other investors (businessinsider), it seems unlikely that the venture is going to make a cut any time soon. The idea of hard-wiring technological components to parts of the human body as sensitive as the brain is going to meet a lot of traction from both sides. Customers in the open market will be hesitant if they want to do a brain surgery every five years, which this device seemingly implies. Software application suppliers will feel constrainted by the inflexibility of platforms, essentially being stuck with the equivalent of producing programs for Internet Explorer 6. Or worse yet, risk having the same situation as the whole global banking system has these days, where the software is still using Cobol blueprints from the 1980s (OneZero). Whether Neuralink has answers to this or not, we will find out on August 28th when it’s due to report on its progress (techcrunch).
Do you think I’m wrong about anything here? I’m most certainly am - and will be happy to hear what you think. Leave a comment below (or in any social media where I’m posting this in).