I’ve been reading Tracey Follows’ excellent new book ‘The Future of You’, about identity. (https://www.waterstones.com/book/the-future-of-you/tracey-follows/9781783965458)
I haven’t finished it yet, but it made me re-visit my thinking about future brain links and electronic immortality. My previous thinking has mostly revolved around having tiny inserts in the brain that signal to and from external IT hardware that effectively acts as a brain extension. My work on machine consciousness looked biokleptically at brain architecture for inspiration on how we might achieve it, and in doing so, made me realise that the timing of signals is very important, and for consciousness to have evolved, early conscious organism brains would likely have architectures that allow neural feedback loops with sensing and processing time of the same order as the time for signals to travel around the loop. This remains likely in modern brains in sensory and conscious areas at least (some parts of the brain would not need such architecture. Without such feedback in sensory systems, the internal sensing of sensing that I assume to be the basis of consciousness wouldn’t work. It’s worth noting that such neural network architecture has no need at all for the feed-forward/back propagation used in training digital neural nets and I think it would make a much better solution. Here’s the link if you want to read it:
More recently, when I was updating my thinking on gel computing, I realised that potential processing speed of the suspended processing particles would be high enough to run many parallel machines in the same gel, and that would solve this machine consciousness problem while also allowing many minds (or instances of the same mind) to share the same gel if it is used to build brain extension/electronic immortality IT. I called it Turing Multiplexing. Again for quick reference, here is a link:
These are just a few of the many trees currently not being barked up in the vast forest that is potential AI/brain link/immortality technology. I am increasingly frustrated that the rapidly growing army of AI researchers seems so determined to all focus on the same corner with the same old trees carrying the same bland fruit, ignoring the much richer, much more diverse parts of the forest with the far tastier, more exciting fruit, over-ripe, falling and wasting on the forest floor because nobody can be bothered to pick it.
However, I missed two tricks myself, unforgivably since I’ve walked right past it many times. In order to make a brain link, tiny IT devices need to be inserted into the brain to connect with each neuron and synapse. Musk’s team call their threads of connections neural lace, but really, we need tinier devices that are wireless. Few people would want to suffer an operation to have wires inserted that can only possibly connect to a tiny proportion of the brain. Devices that are sufficiently small could be injected and float through the brain, anchoring to places they are needed. Simple self-organisation tech is all that is needed to map the architecture and create an external brain extension. The first trick I missed is that instead of merely signalling activity between internal and external kit, these devices could be diverse, with some doing sensing, some storage, some signalling, some processing, some chemical activator roles (such as reacting to or fabricating hormones or neurotransmitters). In my blog on ground-up intelligence, I show how these can automatically link together to create local intelligence that can be broadcast where appropriate, or used locally where it isn’t. Link:
This solution could work inside the brain just as effectively as a city centre.
The second trick is more fun. Applying the idea of Turing Multiplexing to these diverse IT particles, and leaving them in situ in the brain among the neurons and synapses, they can add multi-parallel intelligence, memory and sensory capability without any need at all for external links. They don’t need the server farm, the cloud. In fact, it would be more like an inverse cloud, with cloud-like functionality running in local space, with local ground-up intelligence creating a fractal cloud architecture. You could upgrade your brain by large factors without any external IT at all. Some nice functionality falls out right away:
You could speed up the links between different parts of your brain
Local processing can be done in the synthetic IT at highly accelerated speed and then signaled to local brain wetware, giving higher thinking and reaction speed without compromising natural capability or sensation
Turing Multiplexing of the synthetic IT allows many minds to share the same brain.
Instead of the high latency inevitable in using a cloud architecture, a hive mind could share exactly the same physical space, and since each mind would use neurons in the same region, this would give almost zero latency, greatly speeding up hive thinking, thought sharing, and telepathy.
Body-sharing is easily implemented, allowing much closer relationships. See:
Having the IT resident rather than in the cloud means that privacy, security and ownership issues can be sidestepped. Full control and ownership by the host can be assured (subject to terms of purchase of course)
Electronic immortality is built in. At any point, even after death, the full-spectrum information contents of the entire brain can be read, copied, backed up, duplicated or transferred. Some, such as electrical activity or chemical information, e.g. the state of hormones or neurotransmitters would degrade quickly, while other information such as memory would remain until the biological brain material degrades. But since brain death could be detected, all relevant information can be captured at that moment before it degrades, allowing much longer period for transfer. Nothing needs to be done during life, so it wouldn’t interfere with normal living.
Copying this information to another person’s brain would be easy if they too are equipped with compatible IT. Their Turing Multiplexing would support the deceased person as just another entity running on the same synthetic IT.
A mind could therefore be inherited much as easily as a photo album, and just like a photo album, copied to as many people as desired. Or it could just be left as an electronic copy on a server somewhere in case some future person wanted it. Or transferred to an android body so that the deceased can carry on electronically living. So the same tech that conveys the electronic immortality also allows minds to be shared (though that is also feasible with other electronic immortality tech).
What is really interesting though is that this allows people to have a high quality link to their ancestors. Many traditions have a special place for ancestors. They form a strong part of tribal identity. This technology would allow a full implementation of an ancestor’s mind in another brain, passed down between generations. They would not be some tribal story passed down by words or videos, but would be actual minds carrying on in a living biological body, most likely of one of their descendants. Tribal ancestry would become real in ways never possible before. This obviously needs the technology to have been present when those ancestors were alive.
So there needs to be a first generation. Ancestor Zero! If this technology becomes feasible later this century, as it should, then some alive today will be among the Ancestor Zero generation.
This idea is not new, only this particular technology implementation is new. For millennia, people have had shamans or mystics presumed to have a special connection to their ancestors, to have inherited special knowledge or talents, sometimes via elaborate rituals of inheritance. More recently, computer games such as Mass Effect Andromeda have built on the idea of some sort of IT that monitors the body and mind and can pass on parts of a mind via an AI assistant (SAM in this case).
However, even in the case of computer games, such inheritance is associated with special powers, unique to special individuals . It could become much more commonplace than that. Almost everyone could have such technology. What will we do with it? Will people inherit their parents’ minds, and those of their earlier ancestors? Will government want to control it? Will people need licenses to have their minds carry on in someone else’s body after their death? Will there be requirements for government monitors or supervisors to run in that Turing Multiplex? How would tribal traditions be implemented? Would only certain people have privileges to run certain ancestors or would anyone be able to download anybody?
As always with any interesting development, far more questions are raised than are answered. But it’s fun.