Musk's Newest Startup is Venturing into a Series of Hard Problems:
Tonight [Tuesday, July 16, 2019], Elon Musk has scheduled an event where he intends to unveil his plans for Neuralink, a startup company he announced back in 2017, then went silent on. If you go to the Neuralink website now, all you'll find is a vague description of its goal to develop an "ultra-high-bandwidth brain-machine interfaces to connect humans and computers." These interfaces have been under development for a while, typically under the monicker of brain-computer interfaces, or BCIs. And, while there have been some notable successes in the academic-research world, there's a notable lack of products on the market.
The slow progress comes, in part, because a successful BCI has to tackle multiple hard problems and, in part, because the regulatory and market conditions are challenging. Ahead of tonight's announcement, we'll take a look at all of these and then see how Musk and the people who advise him have decided to tackle them.
[...] An effective BCI means figuring out how to get the nervous system to communicate with digital hardware. Doing so requires solving three problems, which I'll call reading, coding, and feedback. We'll go through each of these below.
[...] The first step in a BCI is to figure out what the brain is up to, which requires reading neural activity. While there have been some successes doing this non-invasively using functional MRI, this is generally too blunt an instrument. It doesn't have the resolution to pick out what small populations of cells are doing and so can only give a very approximate reading of the brain. As a result, we're forced to go with the alternative: invasive methods, specifically implanting electrodes.
[...] Once we can listen in on nerves, we have to figure out what they're saying. Digital systems expect their data to be in an ordered series of voltage changes. Nerves don't quite work that way. Instead, they send a series of pulses; information is encoded in the frequency, intensity, and duration of these pulse trains, in an extremely analog fashion. While this might seem manageable, there's no single code for the entire brain. A series of pulses coming from the visual centers will mean something completely different from the pulses sent by the hippocampus while it's recalling a memory.
[...] One possible aid in all of this is that we don't necessarily need to get things exactly right. The brain is a remarkably flexible organ, one that can re-learn how to control muscles after having suffered damage from things like a stroke. It's possible that we only need to get the coding reasonably close, and then the brain will adapt to give the BCI the inputs it needs to accomplish a task.
Also at NYT, The Verge, Bloomberg, and TechCrunch.
(Score: 4, Informative) by gringer on Wednesday July 17 2019, @10:39AM (11 children)
Okay, time to wheel out my rant [slashdot.org] once again.
A direct physical connection to the brain is not required for a human-computer interface, and it is one of the last places I would choose to put electrodes, given the harm to my brain that would be caused by eventual damage (e.g. accidentally jamming electrodes in further than they're meant to, infection, contamination). Our body includes of a huge bundle of nerves - we've got them *everywhere* - and they're really good at learning things, regardless of where they are. They might almost be as good as a neural network!
If you feel it necessary to jam electrodes in somewhere, do it in a place that is better able to deal with damage and infection, some place where an array of electrodes being ripped out is not going to leave permanent mental damage.
And another thing... there's no need to try to understand how a particular nerve signal neurons work in order to send or receive signals. Just set up an interface to detect and create voltage fluctuations similar to those sent by the nerves, set up random links to that interface and various functions, and let the massive, efficient neural network we've already got in our bodies do the training and testing.
Ask me about Sequencing DNA in front of Linus Torvalds [youtube.com]
(Score: 4, Informative) by takyon on Wednesday July 17 2019, @11:07AM (3 children)
Well, there's the Neuralink cover story: help out paralyzed patients with a brain-computer interface. Maybe you don't need to put wires in the brain in that case. Maybe you don't really need a computer at all. [soylentnews.org]
Then there's the real mission: connect the human brain directly to computer(s) to create a superintelligent (or at least supercapable) human. In this case, you want every cool ability you've ever seen in the movies. Like the implant/interface should be able to access your vision in real time and turn you into a martial arts or parkour master with zero training, with superhuman reflexes. You should have access to Matrix-style neural VR. You try to recall some fact, look at a complex math problem on a piece of paper, etc. and the answer just comes to you as if Wikidata [wikidata.org] and Wolfram Alpha [wolframalpha.com] are natural extensions of your brain. Obviously, it acts as a Babel fish, using lip reading to help predict words as they are spoken, translate in real time, and replace what you would have heard with the equivalent in your language.
Neuralink ties in with another Musk venture, OpenAI. Apparently, they share the same building [wikipedia.org]. Here is the real purpose of Neuralink, in Musk's own words:
Elon Musk Wants to Create Human-A.I. Link and "Make Everyone Hyper-Smart" [inverse.com]
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 0) by Anonymous Coward on Wednesday July 17 2019, @01:52PM (1 child)
With my physique and fitness level, I'd have lousy return on martial arts or parkour instantly gained knowledge. Skeleton, muscles, nerves, senses and brain are components of integral system which must be trained as whole, or something will break. I would do one trick and drop exhausted, or would break a bone, snap a tendon, spring muscle, lose balance, ... OTOH, if I have able other components of the system, it is very likely that I already have been having these mental abilities involved.
Likewise, if you don't know where to begin your quest nor how to eliminate irrelevant, Wikipedia, Wolfram Alpha, or whatever, will just hose your thoughts down with noise.
My point is: if that is the intent, then it is misplaced and the problem is ill-understood. If this technology is at all possible, it will not deliver on its promise, even though it is given that there would be some gain in studying brain activity more closely.
(Score: 3, Insightful) by takyon on Wednesday July 17 2019, @02:06PM
I'm just providing some examples. You can come up with your own examples if you're up to it. Nobody can accurately guess what would be possible if Neuralink's augmented humans vs. strong AI scenario comes to pass. It's post-Singularity stuff.
Neuralink for quadriplegics and other paralyzed people is just a way to make the research look less like mad science. They get their foot in the door, wires in the skull.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 0) by Anonymous Coward on Wednesday July 17 2019, @02:45PM
Then there's the real mission: Scam money from stupid investors
FTFY
(Score: 2) by Snospar on Wednesday July 17 2019, @03:16PM (1 child)
I tend to agree with you, something non-invasive like a sub-vocal pickup and either an ear piece or HUD for
instantresults. I think we're getting close to good sub-vocal recognition now and filling in the other parts of this are Alexa/Siri/Google Assistant level tech. This would give you the "superhuman" ability to search the internet thus appearing knowledgeable (careful with the source data) and even better you could set reminders so easily it would wow your peers (sub-vocal: "remind me I want a beer when I get to the kitchen" sure beats "Now, what the hell am I doing in the kitchen?").Mind you, rather than praise for these new found skills I can imagine being told "No one likes a smart arse" instead.
Huge thanks to all the Soylent volunteers without whom this community (and this post) would not be possible.
(Score: 2) by ElizabethGreene on Wednesday July 17 2019, @06:14PM
One of my memory devices is to picture myself dragging a note up to the upper right corner of my vision. I have a small working memory, and being able to actually do that and then scroll through those would be a superpower for me.
(Score: 2) by Freeman on Wednesday July 17 2019, @04:31PM
But, then, how would we be able to facilitate the eventual Matrix solution?
Joshua 1:9 "Be strong and of a good courage; be not afraid, neither be thou dismayed: for the Lord thy God is with thee"
(Score: 2) by ElizabethGreene on Wednesday July 17 2019, @06:27PM (3 children)
It's not that we have to have it plug into your brain, but that's the location that gets you the most bang for your buck. To give you a practical example, close your eyes and have someone tap the skin on the top of your forearm. If you are like most people you have a sensory resolution of about an area the size of a quarter. If you wanted to stitch in a BCI to your forearm then you need to cover a big area to get enough nerve "pixels" to do something cool. You have more resolution connecting to nerves in your spine, but they are inside a difficult-to-repair bone conduit. You have fantastic resolution connecting to nerves in your mouth, but then your kit is going to get doused in food several times per day. On your hands you have good resolution, but then the kit is in the way.
There is a risk to developing brainpal or neuretics technology, and we'll need a ground-up rethink of security before I'd plug into the internet with one. Still, the possibilities are pretty awesome.
(Score: 2) by gringer on Thursday July 18 2019, @01:21AM (2 children)
Do the same with the surface of your head, and I expect that it'll be similarly bad. Our head doesn't need fine-scale touch resolution, just like our forearm.
We've got thousands of nerve endings in our hands which are the extension of nerves from our arms. These are used for fine motor control, touch, damage detection, and probably a few other things we don't know about. Should be plenty for a small 8-bit interface (i.e. keyboard), or similar.
Ask me about Sequencing DNA in front of Linus Torvalds [youtube.com]
(Score: 2) by PartTimeZombie on Thursday July 18 2019, @01:29AM (1 child)
The end of your penis has even more nerves than your fingers.
I'm not suggesting anything, just pointing it out. You can draw your own conclusions.
(Score: 2) by gringer on Thursday July 18 2019, @04:59AM
Sure, if you're going for maximum number of usable nerves per unit area, the penis, brain, tongue, or maybe even fingertips would be a reasonable place. But I think that the arm (or maybe back of the hand) has enough, is somewhat practical, and is a lot less of an issue if it gets damaged.
Ask me about Sequencing DNA in front of Linus Torvalds [youtube.com]