Elon Musk’s Neuralink: Great news for Humans, disappointing for Superhumans

Sometimes a far-out vision is required to progress human-needs much closer to home. For example, ambitious plans to colonise Mars will hopefully inspire practical solutions to global warming, enabling us to stay on this planet. We should view Elon Musk’s latest venture in a similar light.

Musk recently announced that his Neuralink start-up is developing a brain-computer interface technology that would involve “implanting tiny brain electrodes that may one day upload and download thoughts”. He has said that the company will initially seek to treat people with disabilities, but that “its goal will be to upend human language using brain interfaces”, and: “There are a bunch of concepts in your head that then your brain has to try to compress into this incredibly low-data rate called speech or typing”.

The responses in the press have ranged from naive euphoria to considered scepticism, so here is a perspective based on scientific evidence:

His initial goal is both admirable and potentially realisable. It brings hope that the technology will restore capabilities to those with disabilities. However, his inspiring vision of using electronics to significantly enhance the speed at which able-bodied humans can communicate, is fundamentally flawed. The reason?: It is based on the idea that our brain’s capabilities are constrained by its connectivity with the world, not by the properties of the brain itself, the belief that we are merely lacking faster input and output connections. This seductive idea assumes that our evolution somehow anticipated our technological future, providing us with an over-specified information processor, in readiness for the information age, despite any increased energy demands.

Nearly seventy years ago, when Claude Shannon first defined information in rigorous mathematical terms, psychologists used this new tool to characterise tasks such as listening to spoken words, reading, typing text, playing random notes on a piano, all in the ubiquitous measure of bits per second. Many were surprised at what they found: Karl Küpfmüller observed that “All the instances in the human organism that take part in processing messages seem to be designed to the upper limit of 50 bits/sec” [1]. More recent evidence suggest the figure may be even lower, between 10 and 20 bits per second [2].

The observation that maximum information rates are similar for so many different types of human interactions (and between individuals) is unlikely to be a coincidence. If, as I suggest, this constraint is due to the limited processing speed of our brain, we would expect the information rate of our various senses and physical skills to have evolved to approach, yet not to exceed our brain’s information processing capabilities (Note that physical skills which require monitoring, are limited by our learning speed).

This learning bottleneck appears to be constrained by the speed at which we can integrate what we sense into our internal model of the world. Of all Earth’s species, we hold the most complex internal world model yet we share similar brain physiology with other intelligent creatures. So counterintuitively, this would suggest that the greater complexity of our model would make this process of integration slower, resulting in a slower learning information rate in humans [3].

This might be surprising, but there is some evidence. Some smart chimpanzees in the Primate Research Institute at Kyoto University in Japan have been trained to rapidly memorise numbers (1 to 19) flashed briefly upon a screen. A seven year old male called Ayumu vastly outperformed a more intelligent human (memory champion) in the same test. What to a chimp is merely an arbitrary symbol (“17”), to us is a number within a counting system to base ten, a much broader context. However, our human slowness to learn from the present moment is more than compensated by our prodigious ability to predict the future from our past.

It is worth remembering that we use the same antiquated brain technology or biology, as the apes used 25 million years ago. Unlike our PCs, there have been no significant upgrades to our processor’s hardware capabilities since humans first became conscious. Our brain has incredible processing capabilities but is not infinitely fast. What has changed through history our development of increasingly effective intellectual software, and our ability to exploit external storage devices such as teachers, books and the Internet.

Elon Musk’s new venture into electronic-brain interfaces will hopefully transform the lives of those with disabilities. However, it will ultimately disappoint those who dream of transhuman cyborgs with superpowers. While Artificial Intelligent machines may evolve along a Moore’s Law trajectory, we humans will just have to make the best of the antiquated brain technology we are born with, finding increasingly clever ways to use its incredible capabilities, (and hopefully finding ways of continuing to make do with the planet where we were born).

Thank you for taking the time to read this. If you found it thought provoking, please click the 💚 below.

If you want to go deeper you could read “Bottleneck — Our human interface with reality”, available from Amazon.

[1] K.Küpfmüller “Message Processing in Humans”. In Steinbuch.K. (Ed.): Handbook of message processing, Springer, pp.1481–1501, (1962).

[2] What do we really know about our human interface? in Medium

[3] Networked Humans — the last bit of the connection we forgot to check Video

Neuralink and the Brain’s Magical Future

Elon Musk says humans must become cyborgs to stay relevant. Is he right?