Page 7
He's invented the USB mouse.
Silicon Valley bad boy Elon Musk's grand plan to build brain-machine interfaces to "achieve a symbiosis with artificial intelligence" is obviously more science fiction than fact at the moment. If you missed the presentation from Musk's latest company Neuralink on Tuesday, here's what the plan is: Build an implant that sticks …
This sounds a bit like insurance against his own mortality on Musk's part.
However he is going to need one or two volunteers just for the proof on concept part and it won't hurt, honest!
I think DARPA has been doing BMI research using non intrusive methods for a while now.
It would have to be an 'all or nothing' situation before I would even consider poking stuff into my brain.
Yes, he wants to live for ever.
I know we know very little about how the brain works and this is insane on several different dimensions all at the same time. However, extrapolating wildly, let's say we could, theoretically, simulate an entire brain as a network of neurons (and all the associated endocrine systems and whatever else that might be involved). This is a huge unknown of course – there may well be processes we don't yet understand and/or possibly even something spooky but as yet undefinable (new physics perhaps) involved in consciousness. Who knows? However, let's just say for the sake of argument that we could simulate a full human brain one day using a very large and very fast (by today's standards) computer – and that such a simulation would at least appear to us on the outside to be conscious – so it could talk and learn and hold opinions about Love Island and all that sort of thing – so at a Turning-test level it would walk like a duck and quack like a duck and so may as well be a duck – or in this case may as well be conscious, at least from our external point of view. OK. So let's say one day before Elon dies we can do this (highly unlikely but give me a break here I'm extrapolating wildly). Elon then has his brain 'scanned' at super resolution by some as yet uninvented technology and a copy of his neural network and a snapshot of his 'current state' is constructed in a simulation and, behold, a potentially immortal in-silico Elon would have been created. But wait. That wouldn't be good enough. It wouldn't actually be Elon. It would just be a copy of Elon. Sure, the immortal simulation of Elon would be available for everyone else to enjoy in perpetuity, but that wouldn't be true for the real original Elon. He would still at some point have to die of old age. His own private consciousness would still at some point have to end and so he would still face the prospect of having to grow up and come to terms with his own mortality. So how do we fix that? Ladies and Gentlemen – I present a continuity argument (I'm sure I can't be the first person to have thought of this though). The idea is that we very slowly replace parts of the real Elon's brain with equivalent simulated parts. On the very first day of this process only a small proportion of Elon's brain has been migrated to the simulation – so Elon's consciousness is still mostly running in the squishy stuff and he is presumably still therefore the real Elon. Just a small amount of his neural workload is being handled by a neural simulation. But if the simulation is functionally equivalent to the initially small original squishy part of his brain that has been replaced then what's the difference? If the simulation behaves in exactly the same way for any given input and feeds back exactly what the original would have fed back in response, then is Elon actually any different than he was the day before? Then, month by month, we migrate more and more of Elon to the simulation, less and less of his processing is done in-squishy, more and more in-silico. But every day, presumably, Elon still feels like Elon. His consciousness then remains somehow 'intact' as it is slowly transferred to the simulation and then, eventually, he is all simulation. This way he might in some sense still be the real Elon rather than a copy, and so his continuous consciousness never faces death (well, until he eventually runs out of money to pay for his maintenance and electricity bills and gets switched off, and his parts get reused for the imortalisation or recreation of someone more voguish). However, to achieve all this it will of course be essential that during the migration process there is a mechanism to connect the remaining parts of squishy Elon-brain to the simulated parts – and it will also be essential that the communication between these parts is two way.
That's why he's doing this.
So, in essence, "The Bicentennial Man" in reverse?
The question being: when does a thing that was human become a machine if we slowly replace bits of it with electro-mechanical prostheses, or "Is this really my axe when it has had five new heads and six new handles?"
If, after a head injury or cancer removal, I replace bits of your brain with stem-cell cultures that slowly become functional brain tissue, are you still "you"? How much replacing can I do before you are not?
If you have an immortal, idesructable, immaterial recording and imaging device - or "soul" - those classes of questions become a lot easier to answer.
When I first came to the current research centre around 1995, there was a guy down the corridor who was experimenting with cycling helmets (well it looked like a modern one) and nanovolt low-noise amplifiers and lots of external skull probe sensors....
it seemed to work, move a fuzzy mouse cursor a bit, certainly lots of promise but my colleague basically disappeared , presumably got a better offer from the 1995 Google/Microsoft/.Mil, I haven't ever dared to look where he is...
These things are tiny compared to a human brain. It works on a mouse because their brains are smaller than a nickel (5¢). Conscious thought is considered to be an emergent property arising from the synchronous activity of a diffuse number of specialised "micro-consciousnesses" distributed or arising across the entire brain. He'd be far better off looking at, say spinal transections, where the actual area of tissue that's useful is on the same scale as his device.
Surely the basic problem is we still don't really know how the brain works. We understand some systems quite well, some more vaguely and the basic way that we form consciousness still seems to be largely guesswork.
Also what is the role of various hormones and neurotransmitters? Do we need a chemical sampling / delivery system in order to be able to fully integrate with our noggins?
Given his obvious inspiration from Iain M Banks' Culture novels - I'm sure he's also working on drug glands though. Just hasn't announced it yet. Much more discreet than getting caught smoking dope on a Skype call...
Also, if I remember right, wasn't a neural lace specifically used as a torture device in one of Banks' books?
We don't know for certain, and we never will because the best we can do is find a generalised model. There are too many individualities to say they all work this way exactly. And every single "brain" is broken in its own unique way, and though most conform to the middle of the bell curve, there are those that fall in the middle of the tails that we call extraordinary or even mentally ill, and there are those at the tails that are too broken to work at all.
However, there has been an enormous amount of work completed over the last 40 years, and we have a much clearer picture. We CAN say that there are micro-consciousnesses specialised and resulting in specific neural circuits and that these are interconnected as a result of, for example, brain lesion studies where people will deny being able to see, yet can catch a ball thrown at them or duck to avoid a branch, or will deny being able to see colour, yet can guess with far better than chance the colour of a stimulus, or deny the existence of visual motion (this is a really weird one to conceptualise), or who are conscious of taste when presented with a shape, or a shape when presented with a sound etc. We know that the loss of a 40Hz synchronised electrical signal that might be mistaken as noise in firing even, or as a bias on the neural potential, causes a loss of consciousness that is reported by subjects experimentally and accidentally or as a result of insult.
The role of general neurotransmitters and hormones is also being explored very, very vigorously, although they are thought to be too slow and diffuse to mediate consciousness. Synaptic neurotransmitters are fast enough to form part of a specific signal path, and indeed that is what they do. Electrical signalling to chemical signalling and back again.
And I believe you are correct about the neural lace. It certainly rings a bell.
Surely the basic problem is we still don't really know how the brain works
Given the myriad of different personality types (even the ones which fall into a loose category of "normal", and discounting those which could be considered as buggy, like psychopath, etc) can we even say that there is even *a* way in which the brain works.
I wonder if it's more likely a very basic platform (analogous to a core operating system, as a micro level) but the things which we recognise as "the brain working" are individual, inherited, DNA-influenced things analogous to programs. Each one includes some variations and mutations which define the individual behaviours, personality types, etc.
For example, you and I are likely to approach solving a given problem in slightly different ways - different processes within different programs...not a single thing approach which can be modeled as an algorithm for how the brain works in a generic sense
Personality types are just the decoration on top. Your brain is doing a lot of other work that you pay barely any attention to; processing the input from your senses (vision is a big one, but others need a lot of heavy lifting too, understanding speech requires processing all the way from basic sounds right up to semantic meaning), working all your motor functions and regulating hormones, integrating different systems together (coordinating sight and motor, or sense and motor for example), then there's the entire area of how memory is formed, organised and retrieved. And all this is happening at once in different areas.
How someone goes about solving a logic problem is very very far removed from the underlying operation of their brain.
"Do we need a chemical sampling / delivery system in order to be able to fully integrate with our noggins?"
Working with far too little information to give a definitive answer, but I would say almost certainly yes. Hormones play a hugely important part in the control of many bodily processes including brain function.
That's why (at the risk of sounding like a tin-foil-hatter) I only eat meat raised in EU - no growth hormones
Given his obvious inspiration from Iain M Banks' Culture novels - I'm sure he's also working on drug glands though
Is it obvious? BMIs, and for that matter "drug glands", are old hat in science fiction, as the article points out.
Here's one data point: The first of Banks' Culture novels was published in 1987. George Alec Effinger's novel When Gravity Fails contains both BMIs and drug glads, and was published in ... 1987.
Probably the most famous use of BMI in popular culture, outside that mass embarrassment The Matrix, is in Gibson's various "cyberpunk" novels, which received attention in the media and popular culture wildly disproportionate to their popularity and innovation. Neuromancer was published in 1984.
Poul Anderson's "Call Me Joe" featured a BMI; it was published in 1957.
And so on.
But who doesn't spend hours thinking about the perfect spatula? It's got to be touch enough to lift stuff, yet flexible enough to allow you to get some good scrapeage done in the corners. It needs to be made of a material that won't bugger up non-stick surfaces. Easy to wash, comfy handle. And one of those knobbly bits on the bandle so you can rest it on a sork surface without getting the end (or the surface) dirty is also handy. This is imporant stuff!
See also perfect stirring spoon, perfect cheese knife, perfect main chopping knife, perfect frying pan, perfect washing up brush...
TRT,
I'm afraid the sainted Douglas Adams got that tragically wrong. I'm happy to agree about the carving knife and breadknife. Authority is a good word for that, I've not come across anything better than my Granny's wood handled breadknife - which may well be older than I am.
But the butter knife is so wrong! I invested £3 in a Tesco's butter knife a couple of weeks ago, and it's changed my life! Amazing thing. I should start a company selling them, then go on Dragon's Den and do all my own adverts (like US car dealerships).
So it's a short blade, very wide, it looks like a slighty curvy palette knife. Except it's got those weird serrations down one side you get on cheese knives - the shit ones that come with cheeseboards as Christmas gifts and barely work. It's quite a thin blade, and butter is softer than cheddar, so it cuts it into thin slices straight from the fridge. Then the width of the blade means you can spread even very cold butter if you're gentle, and makes spreading it when warm even easier.
I'm quite fussy about tools, and it's always satisfying to get something that's just right ergonomically.
I've got silicone oven gloves that have actual thumbs, so you can grip stuff - and using those horrible cloth ones without thumbs is horrible, and actually a bit dangerous (because of the lack of good grip).
We certainly do use wooden knives, specifically using wood from the juniper tree for the knives... but the real thing is that we aren't as barbaric as the rest of the world using knives to cut cheese... We use the famous cheese slicer for this delicate and very precise art of cutting cheese!
These sort of gizmos were popular in old Startrek episodes, where they generally involved "pain, agonising pain" when the implantee didn't obey the AI (big computer built into the temple foundations). I'm sure we can trust big tech companies not to come out with anything like that.
"We really need a "re-imagined" Blakes 7. I'm so hanging out for that!"
I've recently started watching Blake's 7 for the first time (I was slightly too young for it originally) and was wondering when there would be a reboot of the series.
Saying that, I might not watch a reboot; I'd be measuring the new cast against the old, and they're pretty hard to match, let alone improve on.
Gan:
You're never involved, are you Avon? Have you ever cared for anyone?
Vila:
...Except yourself?
Avon:
I have never understood why it should be necessary to become irrational in order to prove that you care. Or indeed why it should be necessary to prove it - *at all*.
[exits]
Vila:
Was that an insult, or did I just miss something?
Cally:
You missed something!
As in this account of a poor, hapless victim
SpaceX are a serious company, doing difficult and serious things. And absolutely spanking the crap out of the competition. They're a whole technological generation ahead of everybody* - including several major governments. Plus they've got more shiny stuff in the pipeline, assuming they can also make that work.
A lot of that is down to Musk's vision. Tesla is more of a mixed bag, but getting to scale in the motor industry is incredibly hard - and getting mass manufacture right is actually something that might sound relatively straightforward, but is actually bloody hard to get right in practise. It's like rocket engineering - there are loads of quite well understood problems to solve, but everything has to work well at the same time or effectively nothing works. Was watching a documentary on the German WWII aero industry last night, and they had a decade (and a major incentive) to get mass manufacture right, and they never managed it.
*With the possible exception of Blue Origin - but they've not been tested by real life yet.
"once they extend these devices into equine use, Horses will be able to bolt their own stable doors.."
Some of our horses have been very good at un-bolting their stable door (had to use a secondary clip) and a stallion at playing with knots in string until it unravels. The incentive isn't there to shut the door after them... Not certain if they have that "who me?" level of sneakyness☺
I'd trust that more than a self-proclaimed "stable genius
Perhaps we should start psychiatric evaluations on potential genius businessmen - seems a sensible precaution to avoiding any future 'Hank Scorpios'
- of course of greater priority should be evaluating our politicians first.
Neurolink differs from most of Musk's other big ideas in on important way. Let me demonstrate what I mean:
When we see electric and self-driving cars in sci-fi movies, it's often depicted as part of a utopian future.
When we see sci-fi movies with space pioneers pushing the boundaries to help expand humanity from Earth, it's also often depicted as a good thing.
But when we see sci-fi where everyone has brain implants with the kinds of features Neurolink is aiming for, it is very often shown as part of a dystopian vision of the future.
Quite what the real future holds for all of these technologies, I honestly don't know. I've got an uneasy feeling about this one which I don't have about anything else Musk is doing.
Also, I note that this is just about the only thing Musk is doing that Jeff Bezos hasn't decided to copy. Given that Bezos comes across as a much more level-headed supervillain than Musk, this may be significant.
Some Parkinson's patients today already have have electrical stimulation chips implanted against their spinal column.
Non invasively, electrodes on the outside of the skull can allow, with practice on the part of the patient, crude movement of a mouse cursor. Could be useful for people with motor neuron disabilities.
One should look at science fiction to consider and discuss trends, but one should also look at what is today's technology.
"The best demonstration of this type of technology lives in science fiction, such as Arthur C Clarke's "brain cap".
Surely the best demonstration is the one that Musk refers to himself, the 'neural lace' was widely used in Ian M. Banks' Culture series and Elon is known to be a fan (I give you the 'Of Course I still Love you' drone ship).
Somebody will need to do this kind of work if we are to achieve anything like that and, you have to admit, Musk doesn't lack ambition.
However, it won't be me. I think doctors will have problems justifying this kind of intrusive surgery for Proof of Concept and the various issues around managing infection and other problems around your 'connection port' seem to me rather more significant that concerns about Amazon et all, advertising to your brain - although that will undoubtedly come.
Infection is a huge issue. Knowing people that have permanent medical implants, the are where the skin is pierced will often get infected - and painfully so. However careful they are. And this is a direct connection into the brain, which is normally partially protected from infection by the bloood-brain barrier.
Also swelling in the body is just annoying. Swelling of the membranes round the brain is called meningitis - and the consequences can often be fatal, even if the cause itself isn't that serious.
But there are people with brain implants. I saw a guy with Parkinsons on telly 15 years ago who'd got electrodes in the relevant sections of his brain. It was on a debate on animal experimentation - and he said that he was in favour of vivisection of higher primates - the most troublesome kind.
Because his treatement was pioneered on live gorillas (so my memory says - shouldn't it be chimps?). Then he reached into his pocket and instantly became this hunched, tense, pained figure with loads of muscles obviously in spasm. Then he slurred out that this treatment was to help his Parkinsons, and slowly inched a finger back to his pocket to turn on his battery powered gizmo again. And there he was, an ordinary but quite fit looking older guy.
It was an all the more dramatic demonstration because I didn't even know that technology existed yet.
Ian M Banks' Culture series isn't a blueprint. It's self indulgent transhumanism.
We don't know how a brain works, or what intelligence is. AI, Machine Learning and Neural Networks in real life are marketing, almost fraudulent names. They are not what most people not expert at programming think they are.
Ian M Banks' Culture series isn't a blueprint. It's self indulgent transhumanism.
It's bloody good, self-indulgent transhumanism though ...
(some of Charlie Stross' work is in a similar vein - I live in hope that he and/or Ken McLeod will reach an agreement with Banksy's estate and write a new Culture novel)
Amazon TV was reported to have picked up Consider Phlebas for. TV series, with the director of Utopia at the helm. Since that announcement last year I've not been able to find any more information about the project.
If you have seen Utopia, you might feel that the directors combination of big themes, violence, humanity and cinematography is a good fit for Banks.
Although it isn't sci-fi, I can recommend the BBC Scotland production of "The Crow Road" - pretty much a who's who of Scottish actors of the time (including a younger Peter Capaldi) and it actually does the book justice.
In fact, if you're not familiar with Banks' non-SF work then "The Crow Road" and "Espedair Street"[*] (made into a radio series on, I believe, Radio 4, a few years ago) are good places to start.
[*] - unusual for a Banks book in that it has a happy ending. Sort of.
Simon Ward,
I remember seeing a trailer for The Crow Road, then hearing nothing of it. One of his best books, in my opinion. I think Espedair Street was my favourite of his non Sci-fi - and I'm not sure if that or Excession is my overall fave. So I'm excited to hear Radio 4 adapted it, and disappointed I managed to miss it. Here's hoping for Radio 4 Extra to come up trumps...
His non sci-fi stuff is a lot more uneven in both tone and quality. He himself split them into "nice" books and "nasty" ones. But there were also the weird ones, like Walking on Glass and The Bridge - that don't quite fit into either category. Or Canal Dreams, which I only read once because I thought it wa pants.
I still think that if I'd read The Wasp Factory first, I'd have never read another. But luckily I started with consider Phlebas and The Crow Road.
I've still not read his final Culture book for some reason. Something I need to correct.
'I've still not read his final Culture book for some reason. Something I need to correct.'
The Hydrogen Sonata, again a strong story, expands upon the Sublime ideas featuring a memorable organic lead character and awesome Culture ship the 'Mistake not.....'
I am truly sad there will be no more.
Agreed - The Hydrogen Sonata is very good, and well worth a read. It follows on from Excession pretty neatly. My introduction to the Culture was 'Consider Phlebas' when a guy staying in our flat at Uni said 'you like sci-fi, give this a whirl ...'. I didn't get into his non-SF stuff until quite a bit later, probably after having heard "Espedair Street" on the radio now that I think about it.
I may be misremembering but I think "The Crow Road" might have got the R4 Book at Bedtime treatment at some point as well.
Re-reading "The Steep Approach To Garbadale" at the moment - for someone who claimed to have had a happy childhood and stable family, Banks wrote the 'dysfunctional Scottish family' amazingly well (see also "The Crow Road" and "Whit")
Yes, read the 'Crow Road' myself after getting hooked on IMB's SF output but couldn't really get into it.
I think what sets the SF/Fantasy genre apart is the 'world building' aspect of the writing. This to me adds a whole literary dimension which I think sets the author apart from run of the mill writers.
I do read 'real world' novels occasionally and once in a while I come across something that impresses me but always find myself coming back to the SF/Fantasy genre and that invariably being the novel I can't put down.
To be fair, I'm presuming he's planning to research that.
I personally think this idea isn't going to be possible in my lifetime, for various reasons. But if he wants to spend his money finding out whether that's correct, then good luck to him. You don't generally find the solutions to problems until you start looking for them.
This is early days for the development of brain machine interfaces, probably at the equivalent level of leechs or amateur vivisectionists for traditional medicine.
History tells us we should expect a lot of people to die or be permanently injured before the technology is stable and reliable, and yet marketing is telling us how wonderful it will be.
I think I may have found the problem.
Hooked to Tesla's Autopilot, these BMIs could provide an emergency stop by detecting the passengers' terror in case of an imminent crash. Provided they're not asleep or engaged in more distracting activities.
"In order to assess how feasible these devices are in the real world, you have to have a basic idea of how the brain works. The mass of jelly protected by our skulls houses some 100 billion neurons that constantly communicate with one another by firing electrochemical signals that can be measured by electrodes as a voltage."
Not to mention that the way that the neurons communicate with each other is modulated by how much they've been firing recently, by a number of different hormones, which in turn are produced based on signals going from other neurons into various glands, and by a whole other bunch of neurons sending signalling data from all over the body, most noticeably (and often overlooked) the gut (Hence why 'gut feeling' is not something to be lightly dismissed).
In other words, 'command-and-control' type of interface eg controlling a prosthetic limb, is difficult enough, even though good progress is being made in this area. Interfacing a process of thought, let alone consciousness* is orders of magnitude more complex.
*Not that we know what consciousness actually IS, anyway
Neuralink plans to stick electrodes into brains using a robotic sewing-machine like device
I'm sure billionnaires will. be queuing up for the chance to have one of Musk's little robotic friends flip up the top of my skull like the lid of a pedal pin and then insert doszens (hundreds? thousands?) of extremely small electrodes on particular neurons using a "sewing-machine like device". Why not? They're apparently queueing up get strapped into Beardie's Virgin Death Mincer ( https://www.flickr.com/photos/norbet/40381299060 ... still makes me chuckle)
I'd take a ride in a spaceship, even with a high chance of snuffing it though. Even though I think it's a bit weird that people climb Everest, which something like a 2% death rate - I'd take the same if SpaceX were offereing. And I think 2% is pretty close to the current score in space travel.
I'm not really interested in a sub-orbital joyride though, you can get that more safely and cheaply on the Vomit Comet.
I think I'll wait and see if the space elevator can be built in my lifetime. I'm 53 but very fit and healthy.
I don't see the attraction of such a ride to be honest but then my inner ears don't like amusement park rides.
Mind you if they want anyone to go drill through the Europan ice and see what life might be under it then sign me up. The radiation levels mean it will likely be a one way trip for humans but I'm a biologist and it looks like the best chance of finding extraterrestrial life in this solar system.
Sit in a Bigelow hotel in orbit? no thanks. Still plenty I haven't seen down here on the ground. I'd rather dive to the bottom of the see in a bathysphere. I'd at least sample the mud and do environmental dna and marvel at the nematode diversity.
"The mass of jelly protected by our skulls houses some 100 billion neurons that constantly communicate with one another by firing electrochemical signals that can be measured by electrodes as a voltage."
Pick any PHB - those SOB's can't have more than a few, and those they have firing with barely the power of a 10 watt light bulb. The only electrodes they need are ones connected to various sensitive parts of their anatomy and set to go off any time they come up with a dumb arse idea. The dumber the idea, the higher the voltage.
Elon - I think I might have a few candidates for you.....
The problem is the electrodes can only get signals from the electric part of the proces the chemical bit will be unavailable and anyway sticking an electrode with glass that can detect a neurotransmitter into a synapse is going to stuff it up, literally.
Back in the day during my degrees synaptic research was into stimulating an axon then very quickly slamming the tissue into an ultra cold block to insta freeze it, fracturing it then sputter coating it* and looking at and counting the number of vesicles the terminal had released.
*some steps not mentioned for brevity.
They're not electrodes, they're tentacles. Love the tentacle!
I'm somewhat Autistic, so I actually have some level of personal interest in neurology and keep generally up to date with developments.
The idea that we are in a position to do any form of neural interfacing is comically laughable. Medical textbooks still highlight cases with serious mental trauma (eg Phineas Gage) and are making educated guesses on what particular parts of the brain do based on that, and other similar cases.
Neurologists are just about able to agree on the 8 major headings of mental function, although there is widespread fundamental disagreement over if particular things are linked (ie; if an increased sense of touch also means an increased sensitivity to heat). There is a rudimentary understanding of what chemical trail processes in the brain exist, although nobody actually knows what they do for certain, but are making evidence based guesses. Neurologists are dimly starting to wonder about the effect of allergies that mean people don't eat certain things, and so might be missing entire lines of chemicals that might have some effect on how the brain functions.
So in short, we know about enough to know that we know fuck all.
And Elon Musk is going to stick in a neural interfaces?
I can only imagine that he's particularly short on cash and so is going for another huge infusion of cash to his businesses from venture capitalists, because about the only thing we know how to do is induce a seizure at the moment unless a few hundred years worth of progress has been made in the last few months that i've missed.
I've always maintained a fair bit of interest in this and I think you're right, it's extremely unlikely to be achievable given what we know about the brain.
We're only beginning to scratch the surface of what it does and how. I mean we still don't really understand how it processes information or even something as fundamental as how memory works and how it encodes data. I mean considering that we've only recently discovered dendrites (tiny fibres that link neurones) are actually actively processing data en route shows how little we know
Even assuming that it's doing something similar to a digital computer is probably a huge logical mistake. A lot of the estimates of brain processing power seem to be way, way off as they're basing it on models of a switching computer when it's something very different to that.
We're beginning to be able to see more of what's going on with the advent of improved instrumentation and things like fMRI (Functional Magnetic Resonance Imaging) but, it's still at an incredibly crude level.
I would say we're at the level of a guy from the mid 19th century with a gas lamp and a screw driver trying to decode a microchip, without any understanding of how it works.
There's great scope for improvement obviously, but it's no where near the level where you could interface with a brain directly.
We've had some success though with things like cochlear implants, but that's actually stimulating end points inside your cochlea - they insert a coil with stimulating electrodes along its length into your cochlea (organ of hearing) and it is stimulating various points along that tract, effectively tickling it with electric charges at the point where tiny hairs would normally be picking up particular audio frequencies. It is not a direct interface with your brain.
There's also been some success recently in reading images from your brain with fMRI .. https://www.technologyreview.com/s/425520/brain-imaging-reveals-what-youre-watching/ but again, it's seemingly just reading the pattern of stimulation of neurones form the "camera feed", rather than how they're being processed. It could still be useful for designing ways of getting optical information into your optical nerve in blind people.
Indeed, the other thing we can do is stick a glass needle into a neuron and release a chemical which fluoresces in proportion to the amount of say calcium ions there are. Then with a microscope of sufficient power you can watch the dendritic tree light up at at least some of the synapses, the Gabbaergic ones at least.
I learnt something from looking stuff up raised in Peter Tse's book: muscarinic acetylcholine receptors are g-protein coupled, not ion channels. So they work by neuromodulation. I don't think that was known back when I did my degrees. But I only dealt with nicotinic receptors on muscle. I labelled them with 125I or rhodomine labelled alpha bungarotoxin. I used the beta bungarotoxin to denervate mouse embryo limbs, it's a phospholipase with a neural specificity, chews them up from the synapse back.
This is exactly the reason why I propose a non-invasive method of external fixation of brain activity, as the creation of personal profiles (AIs). That is, if the profile can adequately answer all the questions, then there is no sense to go into the internal mechanics of the brain, better to leave it to neurosurgeons and begin to do science.
I mean my Lexical Clones, like clones of Plato, Presidents Roosevelt, Clinton and G.Bush, Julius Cesar, Bernard Shaw and Fedor Dostoevsky. Their texts contain traces of their internal neural networks as hierarchies of patterns. I spoke to them, read my report at NIST TREC 2003-6?
This is exactly the reason why I propose a non-invasive method of external fixation of brain activity, as the creation of personal profiles (AIs). That is, if the profile can adequately answer all the questions, then there is no sense to go into the internal mechanics of the brain, better to leave it to neurosurgeons and begin to do science.
But don't we already know that many if not most of the brains activity is being passed around by chemical reactions? If so, then why the focus totally on the brains electrical signals? Electrodes on their own aren't going to be capable of reading most of the information flowing around them.
I do have more than a passing idea of how the brain works. I understand a lot of neurophysiology, I've worked with people sticking old style drawn glass electrodes into brains and muscles etc. Neuronal function is much, much more than the electrics, the whole point of making neurons communicate via chemical signals at synapses is that each synapse is on the fly reconfigurable (in theory and practice) which will change the receiving cell's sensitivity to other incoming signals. Also we know the cell bodies can send retrograde action potentials up into the dentritic tree to change the settings of the synapses.
In addition there is evidence and a plausible mechanism that some pairs and groups of synapses on dentrites can act as logic gates, which means you can get computation at dendrites.
If anyone wants to get on a steep neurophysiology learning curve with added formal logic check out Peter Ulric Tse's The Neural Basis of Free Will. But be warned I have a Physiology PhD with a neurophys background and I needed to looks stuff up to both update myself and remind myself. Others may need a physiology textbook at hand, and the internet.
Those wires btw are nowhere near fine enough to record from dentrites let alone individual synapses, they are only getting the big neural action potential spikes which are just the summed output, if you haven't measured ALL the input and its effect on the receiving cell you won't know or understand squat. Neurons can have huge dendrite trees with thousands of inputs.
One thing to note, the inhibitory inputs are clustered on the cell body or actually on the axon hillock so if nothing else we have the free will of an editor, the ability to say NO! After all part of what we regard as mental illness is the inability to do this. To feel you have to act on every thought.
I am not a neuroscientist, but I worked for years developing software in a neuroscience lab, and everything that you've said here is entirely consistent with what I learned and observed in that role.
"Those wires btw are nowhere near fine enough to record from dentrites let alone individual synapses, they are only getting the big neural action potential spikes which are just the summed output"
Thanks for bringing back many sleepless nights! This is exactly correct, but with multiple electrodes and some fairly hefty processing power, you can make some magic here. We were able to isolate and analyze individual synaptic behavior with the proper setup.
But that sort of thing wouldn't really be possible with what Musk is talking about.
When first deployed into a volunteer, Mr Albert Stoat, the approach showed some drawbacks viz:
1) Post operation, the subject would only speak in bursts, and then only to say "Aggleaggleaggle".
2) It turned out that during post-operative recovery a malicious person or persons unknown had installed a coin miner in Mr Stoat's firmware.
3) An MRA scan of the patient's brain later showed activity in the hypothalamus, which turned out to have somehow been overwritten with obscene MPEG shorts.
4) When the diagnostic interface was enabled and connected to a terminal, it would only play animated cat gifs.
5) Upon applying EST the patient became lucid, although only to recite a list of some four million, three hundred and ninety-two thousand, eight hundred and four userid/password pairs, after which he said "Aggleaggleaggle" and passed out.
6) Mr Stoat reacted poorly to static electricity shocks, passing out then immediately waking, saying "Aggleaggleaggle".
This is how he plans to have autonomous cars by next year. He's going to offer these implants for "free" in exchange for an hour or two of slave labor per day driving an "autonomous" Tesla as a robotaxi. Everyone wins, at least until the implant becomes sentient and Skynet is born.
Recent neurological research says
a] The brain have been discovered to be communicating within themselves via brainwaves to keep parts in sync, it's those often monitored via EEG.
b] The neurons actually fire off differently in different situations, responding to a greater number of stimuli than the single one previously thought to trigger them.
This will make any brain connection or cyber jacking much more complex to decipher and slow the development down greatly or limit the functionality to selective diagnostics or basic prosthetic function only, not really the scifi brain connection of the future yet.
I'm not sure that the high impedance of the fibers are necessarily a bad thing. I'm assuming that bone isn't a good insulator with respect to RF so the impedance can be used with a trans-impedance amplifier to get decent SNR for the signals from neurons but what I wonder if this is even useful if we have no idea what individual or small groups of neural activity (synaptic activity, right?) means in terms of function. I know that we've identified what parts of our brain are in terms of function but I don't know if we can make a map of that plus the individual neural activity to give any meaning to the data collected. It all smells like a PR stunt to me or maybe a grant letter but for private investors.
"After learning this information, these BMIs should be able to do the reverse and stimulate certain areas in order to kickstart certain brain functions. "
Anyone getting this done, based on this statement, should lose the right to vote or hold any public office, because this just said that it could be used to reprogram your brain to think how someone else wants you to think.
Corporations would love this though. Imagine... I've GOT TO HAVE A TESLA NOW!!! 200K? WHO CARES!!! GIMME GIMME GIMME!!!