Elon musk

Specially speaking at an AI autonomous vehicle industry event this week and was asked whether there is any kind of connection between Elon Musk’s efforts with Tesla and his activities with Neuralink, which is his futuristic play on using advanced tech to tap into human brains.

The question was prompted by last Tuesday’s Neuralink presentation (on July 16, 2019, see the recorded video here), showcasing Musk and various Neuralink scientists and researchers, touting their newly unveiled wares and software that they’ve been toiling away on, having crafted it with great intensity and single-minded purposefulness in their souped-up neurotechnology labs.

My answer about the potential connecting tissue between Tesla and Neuralink is that the simplistic answer is there isn’t any direct linkage per se (other than Musk himself), but, if you are willing to stretch your imagination, stepping outside most anyone’s comfort zone, there might be a synergistic aspect in the realm of self-driving cars.

Yes, I said it, there is a potential connecting perspective via the realm of self-driving cars.

Preposterous, some might retort.

Well, give me a moment to at least explain the logical underpinnings that support such a seemingly outlandish assertion.

Let’s start by unpacking what Neuralink is doing.

Unlocking The Human Brain

Neuralink was one of those stealth firms that Musk opted to get underway, doing so in 2016, becoming publicly known later on in 2017, and gained notoriety due to the hiring of several of the world’s topmost neuroscientists. You can’t really keep a brain trust of this magnitude under wraps for very long (brainiacs about the brain, one might say).

On a practical everyday basis, the stated goal was to be able to better treat brain diseases and that via neurotechnology you could hopefully plant devices into the noggin and overcome or at least mitigate neural disorders.

Like many newborn “hot” tech firms that want to also have a broader vision and see far beyond the near-term trials and tribulations of today’s world, they also stated that they were ultimately aiming to enhance humans, essentially bringing to real-life the desires of transhumanism (a philosophy that high-tech should be used to transform humans to a greater plain).

For start-ups that want to swing for the fences, it’s crucial to have an out-there kind of aspiration or goal, and undoubtedly Neuralink has one.

Imagine that you could implant electrodes or the equivalent into the human brain.

Suppose further that you could do this without harming the brain, and you could do this without having any visible indication that someone has these in their brain (no protruding antennas, no Frankenstein scars, etc.).

To use the implanted electrodes, there would be a wireless means to communicate with them. Thus, you could have these embedded computer-like elements sitting inside your head, no one else the wiser, and yet you would potentially be the wiser, or maybe at least be better off, as a result of the brain melded apparatus.

This is a really tough problem to solve.

Consider these foundational questions:

• How can you put something into a person’s head and yet not kill them or otherwise mar their brain?

• Even if you can get electrodes into the brain, what good will it do if you are unable to detect what is happening inside the brain?

• And, even if you can plant something and even if you can perhaps detect signals of one kind or another, how can you interpret and make sense of those signals such that it relates to what people are actually thinking?

I realize it seems eerie to contemplate putting electronics into someone’s head.

You’ve perhaps seen pictures or videos of various methods and devices being used today, mainly consisting of overly sized, physically intrusive electrodes that stick-out like a porcupine and you can only viably utilize while inside a sterile hospital room or equivalent.

There are some efforts involved in “remotely” detecting the neural signals of the brain by wearing a cap or bowl-like bonnet on your head. These though are oftentimes unable to pinpoint specific neurons and also have to deal with signal degradation and noisiness because of the cranium or skull acting as a kind of ethereal barrier to signal detection.

The initial joke or eye-wink within Neuralink was that Musk wanted them to build a “wizard hat” for the brain.

Zoom forward in time to the Tuesday presentation and you can see that the researchers have gone further than merely aiming at devising a hat.

They have been working on tiny electrodes that can be inserted with the use of tiny threads, doing so with delicate precision as guided by a specially built robotic hand. This also includes an electronic chip that is called the N1 sensor and a wearable device that looks like a petite hearing-aid that you might clip onto your ear (they call it the Link, a portable Bluetooth device that includes a battery).

I’d urge interested readers to look at the Neuralink white paper that describes the particulars of the clever approaches used and includes pictures and diagrams depicting the pioneering devices involved (here’s a link to the white paper).

As stated in their white paper, they have been developing a groundbreaking, scalable, high-bandwidth brain tapping system, encompassing flexible polymer probes, various customized low-power electronics, and can be precisely placed or inserted via the use of an innovative robotic hand. Besides having put together the devices and infrastructure to do this, they have also been running experiments on rats (in accordance with the lab animal guidelines of the National Research Council).

Believe it or not, they have created a smartphone app that you can then use to communicate with the devices inside your head. Instead of watching those endless cat videos, you can watch your brain at work, though it is right now just measuring bio-electrical spikes and signals. Figuring out what those signals signify is still being figured out.

Before you rush to your local store to buy electrodes, please keep in mind that this is an early-stage research effort and they have quite a distance to go before this becomes something tangible in terms of life-changing for humans. Some critics have also pointed out that their efforts do not seem as yet to involve the traditions of peer review and associated experimental control trial methods, plus they have not sought FDA approvals and nor other regulatory review and feedback.

On a scoping aspect, it might be worthy to note that the brain has an estimated 100 billion neurons (some argue it is closer to 86 billion, which admittedly is nearly 20% less than 100 billion, but anyway, let’s not quibble over it). For each neuron, there are biological synapses that are like branches outreaching to other neurons, of which there are an estimated 100 trillion of those.

The Neuralink team suggests that they could potentially insert up to 10,000 electrodes into a brain, a huge leap forward from today’s technology, yet if you compare this amazing step-up to the vastness of the brain, you obviously are still relegated to a teeny tiny portion of the universe of neurons in your head.

Musk, always wanting to move faster and make jaw-dropping rapid progress, stated during the presentation that Neuralink might be undertaking real-world trials in a human head, an actual living human patient, before the end of 2020.

He didn’t say how to contact the lab if you want to volunteer, so if you were thinking about it, you’ll likely need to wait a bit longer than 2020 to try one out.

Brain Insertion And Self-Driving Cars

In an AI column that I posted nearly two years ago, I had speculated about the progress that we’d see in the field of brain tapping or more properly and formerly referred to as BMI (Brain Machine Interfaces). You might know of this via its slang or streetwise naming, the so-called "brainjacking" as a more commonplace vernacular.

It is wonderful to witness the BMI progress purportedly being made via the work of Neuralink (subject to verification and validation).

With that being said, let’s return to my earlier remark that somehow BMI or the capability of tapping into the brain could be leveraged toward the advent of self-driving cars, a likely puzzling comment of intriguing proportions.

I’ll start by offering an important utterance by Musk during the Neuralink presentation, whereby he said that the effort to connect with the brain is going to be a type of symbiosis with Artificial Intelligence. He even said that it would be a “go along for the ride” of the brain and AI being able to work together, a virtual merging of a kind, combining what an AI system brings to the table with what the human mind brings to the table.

Couldn’t agree more.

The current approach by Neuralink concentrates on being able to detect neural signals in a human’s brain and then allow the human to do something about it.

You might for example glance at your smartphone, see that it is displaying the spiking of neurons that have to do with (let’s say) being depressed, giving you a handy visual indicator that you are on the verge of a potentially deep and debilitating depression cycle. By having this detection so easily displayed, you might re-focus your mind, aiming to halt the depression in its tracks, doing so before it can grab hold of your brain overall.

Up the ante. If there was a sophisticated AI app on your smartphone, it might be able to interpret those neural signals just as you would have, likely faster, and perhaps alert you that a depression cycle was festering.

Or, and here’s a bit of a controversial next step, the AI might opt to send signals to the electrodes in your head, trying to suppress the depression. This could be done by the AI without having to warn you about the emerging depression. Instead, the AI goes about doing its job, just as though you were using an app to keep you from forgetting where you put your car keys or neglecting to brush your teeth.

Whoa, some say! We have gone from the read-only version of brain tapping and ventured into the write-into the brain variation, a nightmare of sorts to those that doubt the efficacy of such an approach. Couldn’t the government then force our minds to be filled with the thoughts that the government alone wanted us to believe? Yes, it opens an ethical can-of-worms, for sure.

That’s why the word “brainjacking” is befitting when you are allowing a BMI to cause a change in the neuron signals within the brain. We probably would agree that jacking isn’t the right wording if the BMI only was reading from the brain. Once the system is permitted to change the brain by altering signals in neurons, you’ve entered into an entirely different ballgame and essentially can overtake or jack the mind, presumably, theoretically.

Anyway, I’d like to make this discussion easier by stipulating for the moment that we’ll assume the BMI will be read-only (conspiracy theorists won’t like the assumption).

Now, let’s bring self-driving cars into this picture.

Currently, you drive a car by handling a steering wheel and moving your feet onto and off of an accelerator pedal and a brake pedal. That’s the long-time conventional method of manipulating the driving controls.

Some have suggested that with the advent of Siri and Alexa, considered Natural Language Processing (NLP) AI-based systems, maybe humans could merely speak while sitting inside a car and use spoken commands to drive the vehicle. Hey, driving system, please accelerate to 40 miles per hour. Driving system, you are going too fast, slow down to 30 miles per hour. Hey, hits the brakes, hard! And so on.

Not at all practical.

The time it takes to speak such utterances tends to be much lengthier than moving your arms or legs, thus it would be too much of a delay. Your utterances might be misstated, especially when faced with a dire emergency. You also might say the wrong thing, perhaps telling a friend in the car about the time that you hit the brakes, and the voice system mistakenly suddenly jams on the brakes since it assumed you had given it a command. Etc.

What is much faster than voice and even faster than using your arms and legs? Your mind.

Let’s suppose you had a BMI device in your head that could connect to your car, doing so wirelessly, just as your smartphone can today connect wirelessly to your car. You get into your car and mentally tell the car to start. It starts. You mentally tell the car to proceed ahead at a pace of 5 miles per hour. The car proceeds ahead.

Before you howl at this seemingly crazy idea, I concede that there are numerous problems or issues with the BMI-to-car approach.

Even if you could master your mental processes to the degree of being able to focus them toward driving the car and doing so in a manner that the BMI can detect, you are undoubtedly skeptical since a person might veer away from the mental effort of driving and suddenly the car has no mental orders being fed to it. Or, maybe the human “driver” is drunk and their thoughts are wild, causing the car to wildly weave across the lanes of the freeway.

Well, remember the symbiosis point.

If the car was AI-equipped, it would be co-sharing the driving with you, and presumably could contend with situations whereby your mind wonders afield or you are in a drunken state-of-mind. Almost like having a sober buddy to help with the driving.

Usually, at this juncture of such a discussion, I have someone say that you might as well have the AI do all of the driving and cutout the human driver. Why deal with the BMI aspects if you don’t have to do so, some would argue.

That’s the rub.

We don’t yet know whether we can get AI to the vaunted Level 5 of being able to fully autonomously drive a car. Some say yes, we can, and others say we aren’t going to get there, either for a very long time or possibly never.

As such, if you believe that humans will still need to be involved in the driving task and that the AI can’t make it the “last mile” of being good enough to do so, we might augment the AI with the human brain, using BMI. For example, it’s rather apparent right now that we aren’t going to crack-the-code anytime soon on human common-sense reasoning, therefore you might allow the AI that doesn’t have common-sense reasoning to tap into a human that does have it.

Conclusion

Before you bust a gasket and point out that this have lots of unsightly loopholes and guffaws as an approach to driving, I’ll just say that I had forewarned you that this is outside-the-box thinking. Maybe outside of the box of the outside-of-the-box thinking.

It’s an interesting equation, offering that we might combine BMI with AI, and then combine those with cars, leading us toward either better semi-autonomous cars or fully autonomous ones, though obviously, it isn’t in the spirit of being truly fully autonomous if the AI is going to be reliant on the human mind for assistance.

There’s an argument in there about the notion that supposes your brain was able to interact with the AI of the self-driving car on a subliminal basis, thus, you weren’t really taxing your mind per se and it was like you were merely along for the ride (meanwhile your brain was being used to help out).

Does that constitute full autonomy, when you aren’t even aware of your brain being so utilized? Most would say that if the mind is used in the tiniest iota, the AI is not truly autonomous.

As relief on this topic, none of this is going to be possible for quite a while, so you can allow your mind to rest easy about this, for the time being.