You Musk be joking: A mind-reading Neuralink chip in a pig's brain? Downloadable memories? Telepathy? Watch and judge for yourself
- Reference: 1598662349
- News link: https://www.theregister.co.uk/2020/08/29/neuralink_pig_brain_interface/
- Source link:
The porcine test subjects were: a control animal; one who has had the implant in place for a couple of months; and another who had had one removed and looked healthy, presumably to reassure people who want to eventually wire a Neuralink gadget into their gray matter that the process is reversible.
As Gertrude, the pig with the coin-sized implant, walked around, ate things, broke wind, and did other things pigs do, what was said to be her brain activity appeared on a screen, revealing real-time neural network activity observed by Neuralink's embedded tech.
Musk said the implants have been performing well, and his eventual goal was to build human implants that would enable the sharing of memories, telepathic summoning of Tesla cars, hands-free gaming, and so on.
Essentially, the plan is: you think a command, it's detected as electrical signals by the Neuralink electronics fitted in your brain, and then wirelessly passed to a computer or some other device to interpret and obey. The system could also read your mind, help you overcome brain injuries, etc.
"You could store your memory as a backup and restore the memories or download them into a new body or a robot body. The future's going to be weird," Musk said in a web broadcast, embedded below. He added it was like having a "Fitbit in your skull."
[2]Youtube Video
Version two of Neuralink's automated implant installer was also shown off at the event, a system that can install the mind-reading widget's 1,024 five micron-wide electrodes 43 millimetres into the brain to scan for electrical activity. The surgery robot, seen below, was designed by an outfit called Woke Studios for Neuralink, and it's hoped that the Neuralink "Link" can be embedded in your head without the need for general anesthetic.
As the device design improves, the team wants to go deeper, if it's safe, to increase the sophistication of the tech.
[3]
The robot designed to surgically implant the device into brains ... Source: Woke Studios. Click to enlarge
"The first builds are going to be quite expensive but that price will rapidly drop," Musk promised. "Inclusive of surgery the cost will come down to a few thousand dollars, the same as Lasik surgery."
Neuroscientist used brainhack. It's super effective! Oh, and disturbingly easy [4]READ MORE
He said spoken words have a very slow data rate, and speech is so very poor at communicating compared to analyzing and interpreting raw electrical data straight from the brain. Thus, if we can communicate uncompressed thoughts at high speed to others wirelessly, this would bring about something close to telepathy. Senses could also be expanded using hardware, to give "super vision," he suggested.
The inclusion of pigs raised some questions. Musk explained that pigs are a good substitute for humans, with a similar dura membrane and skull shape, and will have to do until trials can begin on humanity. "It's easy to keep pigs happy, they love food, it's a true thing about pigs," he said. "Give them some food and some friends and they're happy. Pigs are quite similar to people."
Other animals have been tested with the Neuralink tech, we're told, with roughly nine in ten implantations succeeding.
The initial focus of the four-year-old company, largely funded by Musk, is to help people with disabilities regain some control of their bodies, the billionaire said. What's the betting we'll see this before or after the arrival of power-generating fusion reactors, viable quantum computers, or faster-than-light travel? ®
Get our [5]Tech Resources
[1] https://www.theregister.com/2019/07/18/musk_neuralink_brain_interface/
[2] https://www.youtube.com/watch?v=DVvmgjBL74w&feature=youtu.be
[3] https://regmedia.co.uk/2020/08/28/handout_woke_neuralink_robot.jpg
[4] https://www.theregister.com/2019/08/07/brain_hacking_bsides/
[5] https://whitepapers.theregister.com/
Re: Eyes Back Head
"So, with a webcam or two, Musk could have eyes on the back of your head."
FTFY
Reductionist science, thus fatally flawed speculative tech.
Every person's brain, nerves, senses, and muscle structure are probably unique, probably even for genetically identical clones too, so it'll probably always be a significant customised challenge to encode/decode neural signals. The bodies communications are probably rather more complex than realised, using electrical, radio, and chemical signalling.
Remote control, not using muscle nerves, will probably be a lot harder than they suggest; even prosthetic limbs using muscle nerve signals probably require plenty of user training to learn to use, because of the above uniqueness. Complex senses like sight and hearing are probably going to be damned hard to do if attempts are made to bypass the learned, complex structure and circuitry of the eyes and ears.
Backing up memories is not happening with so little and only electrical sensing, you'd need something like a 3D "Culture" "Lace", grown into the brain, to do that, and probably outside the skull too e.g. some of what enables us to think maybe outside the skull too, like neurons in the spine and gut, and several semiconducting radio nodes (meridians) spread across the body.
Some people have suggested that the brain is a mass of continuously competing non-linear entities, with the consciousness providing the illusion of unity; hardware simulation of neurons will probably never create a true AI, including so-called Quantum computers, which probably won't scale.
...it won't be the one implant, will it? It'll be tens of thousands of implants which will be as easy to remove as his blasted space vandalism satellites
Elon-gated kit
Musk tends to dream big dreams and build workable kit as a step towards it. But the dreams don't go away and often just get more outlandish. Like his self driving Teslas which allow you to watch movies instead of drive. Right up until you smash into a parked police car (in the latest incident). And while SpaceX rockets and capsules are great, colonizing Mars is, at best, problematic and terraforming Mars is crazy.
This latest dream could be a useful development in automated implant surgery and even the neural link could benefit certain severe cases of brain malfunction.But I wouldn't want to see the rest of his pipedream happen even if it could.
Optimistically...
I'm thinking about when I write code - would my thoughts go directly into the program I am writing or perhaps direct to my fingers? How about when designing systems - would thoughts go direct to diagrams or flow charts? Then there is database design and all the relationships.
Oh the possibilities...
Re: Optimistically...
"would my thoughts go directly into the program I am writing"
No, they will go directly to the mothership for processing, and only those thoughts deemed proper and correct will be returned to you for action. But first a word from our sponsor ...
Re: Optimistically...
You're believing his bullshit if you think it will work like that. Instead it will be a choice of typing at 80 wpm or whatever you normally type at, and maybe 8 wpm if you're lucky, with the brain interface. All he can do is pick up electrical signals, it can't read thoughts or even whole words out of your brain. You have to think about each individual letter.
This is more Musk hype without substance, like his claims for full self driving Teslas before the end of 2016 2020.
This might be a great thing for people who are paralyzed, assuming it can be faster than the eye tracking they currently use. We are decades at best away from it being faster than a manual process like typing. We will not live to see typing by thought faster than a keyboard for the average person.
Shiny happy neurons everywhere...
If you spend time working with the predictive engine people on a project, you gain a lot of insight into the limitations that multiplying large arrays can offer those seeking quasi intelligent results.
A speaker at GopherCon 18, talked about a system that you jammed go programs into, and it got upset if your source was unusual.
He gracioiusly showed the early efforts, where they fed the array a bunch of good go programs; then fed it new programs. The prediction engine became upset, because the copyrights in the comments didn't match those of good programs.
Try 2, feed the output of the a lexical analyzer into the prediction engine; better, but unless you follow Pike/Cox naming convention, you are a radical outlier, thus suspect.
Eventually, feeding the abstract syntax trees yielded more useful results; possibly an automated code reviewer, that doesn't judge, just states : "that is unusual".
If not for Colorado's lenient drug laws, I might be able to recall the speakers name, and more details of the talk. He may have spoken French, but I was kinda high.
At any rate, it did illuminate the possible value of unconstrained search in a restricted domain, in a way that unregulated markieting bullshit from companies like TESLA never do.
OK, so we can extract thoughts or perceptions from an animal...
How long until some bright spark figures out how to put thoughts and perceptions into an animal's brain? And what happens to freedom of thought and individual identity when that starts getting rolled out to people?
Oink!
> "Give them some food and some friends and they're happy. Pigs are quite similar to people."
Apart from the being happy part.
Re: Oink!
So true. Most people are nowhere near as happy as my pigs.
Happy hogs are tasty hogs.
Eyes Back Head
So, with a webcam or two, I could have eyes on the back of my head?