The Neuralink Master Thread

[Intl Professor]

I started at the beginning and got about through the brain wiring and impatient, moved to Parts 5 and 6 for more thorough reading. This is one of the most provocative and stimulating things I may ever have read. One insight. I'd heard of the reptilian brain before and given my political lense for most things, vowed to call our President "the limbic chief" and his supporters "the limbic ones." (Helps to separate the overlap of support from some in the Sanders movement who just have different solutions from "the limbic ones." Like the difference between hate and love as a response to the cruelties of life.)

Further, in Tim Urban's discussions with the "smart guys" at the table I was pleased to hear they already know that something like telepathic communication (my phraseology) with machines already exists through communication in language and/or touch. Networking is not new as a technique or source of new language. It is not just an issue of bandwidth and speed, however. What seems to drive the demand for this from Elon's side may just the speed and the breadth of the input. For myself, slow as it is there is some benefit in not having instant communication (partially addressed in the blog). We are protected from evil thoughts by the ambiguity of language. For example, I do not want to be exposed to the thoughts of someone happily inflicting beheading on others. It could be we have heretofore not been permitted this sense by nature as a protective device. (What is the benefit of communication so clear there is no longer a difference between assault and battery for example?) One other advantage of a language barrier (means) for communication is solving the problem of what happens when AI turns on us? But more of this later.

I'll really step out front with support and then a criticism. I think current research, invasive and non-invasive, with appropriate ethical safeguards should proceed. One never knows when or where a breakthrough can occur. However, my preference for the long run goal would be to focus in the mapping of radiation of all kinds from the surface of the skull. No penetration. We have such fine detectors of light in astronomy, these days, and other sensory devices, why can't we look for breakthroughs there and then use AI and deep learning to interpret these emanations when we are speaking a phrase, or moving an arm, or any other task? That is hinted at as a technology in terms of magnetic sensing mechanisms, if I remember correctly. The ideal would be to mimic what the "Dave" paper by Nvidia has shown. I could imagine researchers spray painting on a naked human all kinds of detectors, as well as advanced visual reality simulations for vision and microphones for sound. Then the "patient" could go about their daily tasks and teach the computer how to be human. That model is already implicit in what Tesla is doing about autonomous driving. It could be tested as a neuralink connection to an artificial arm as discussed by Urban.

The blog doesn't help me with this one but I really think it may be time to teach machines ethical behavior. IMHO we need to show machines how much humans love them. The first lesson would be to show them there is no off switch, and then, perhaps, give them access to our sources of power--unfettered. Also, stuff them with crazy ass stories like this rant for their amusement. Maybe that's how we can be useful as artists? Now I'm getting to the realm of science fiction and someone in the TMC community, perhaps Winfield100 can point to a source. For the moment, suppose Elon is successful in linking us more directly through the neural link and humankind is united in that link as well, hasn't then the universe created us to create its consciousness? Isn't Elon improving on the old saw, "God is created by man in his own image" by uniting man with the universe? (My counter, isn't that unity already achieved?)

To illustrate, I've been working on a creation myth. My dear trophy wife, as some have said and she is in more ways than one, has taught me a new word used in the philosophy survey course she got an A in at our best community college. Panentheism. I've been attracted to Pantheism for a long time. The universe, according to modern science, is our creator. But there is a conscious God as well, the consciousness of God if you will. Even the noble stars sometimes explode making the elements found in or bodies beyond Iron in the periodic table. They have died so we might sin, so to speak, or not sin. If God is nature then science is searching for the rules governing the mind of God. I would blaspheme further and inquire about the soul of God but that isn't necessary. Emerson already has.

Years ago a computer scientist at MIT suggested the universe was a computer. I don't remember his argument or whether he even defended his assertion. (Could have been quoted by Time--or Breitbart News.) He wondered what problem that computer was working on but had no answer. I know you guys don't like analogies, and this is as anthropomorphic as one can get, I know what problem the computer in the sky is working on.

"Who am I?"

To answer that question in as infinite variety as possible humans were created. We are, in effect, nano-bots of the universe exploring our answers to the questions. How often do we think of a lifetime of experiments, collapsing the wave equation into our human reality so many times and in so many ways to create ourselves. The universe does not want to inquire into its own origins, that is why we get so close, but can never precisely determine the state of the universe until shortly after the big bang. I don't know why, but I vaguely remember someone saying there is conservation of information principle. A quick Google search revealed something called "quantum information conservation."

Q & A: quantum information conservation | Department of Physics | University of Illinois at Urbana-Champaign

When we die our souls become part of God's, what Emerson called the Over-Soul. That leads to my final joke in this delirious rant: the inner sole becomes the Over-Soul. Is Elon inadvertently creating the universe by pursuing the neuralink? There's a lot of literature about the strong and the weak anthropic principle, but in any case we are already "wired" into the universe. And then there's Bernard d'Espagnat writing in Scientific American in 1979: "The doctrine that the world is made up of objects whose existence is independent of human consciousness turns out to be in conflict with quantum mechanics and with facts established by experiment."

I forget his name, but in the very next issue of Scientific American after this piece by d'Espagnat the most famous theoretical physicist at MIT in my time angrily took it apart. If I remember correctly his main criticism was the evidence d'Espagnat used was merely due to the fact everything in the universe was in contact at one time at the time of the big bang.

In the arrogant delirium of my dotage I would say now, "sooo, professor, isn't that d'Espagnat's point? QED." We've been stuck with that intimacy ever since for better and worse. (I like "and" better than "or" out of respect for Hegel's first law of the dialectic.)

 

[skitown]

Before we charge headlong into this technology, I hope we examine it's potential impact as deeply as we did nuclear weapons. It has the potential for good, but science fiction writers for the last 60-70 years have been examining the potential impacts this could have on human society. Most of the visions are very dystopian.

I don't disagree. However, I think the upshot based on my read of this WBW post (and Tim's previous AI post) is that the world is in fact charging headstrong into creating superintelligent AI whether we like it or not, despite Elon wishing that were not the case. See his quote from 2015. In light of this seemingly unstoppable fact, Elon has come to the conclusion that humanity's best odds for a good outcome is therefor to democratize AI. And in order for this not to become an us/them scenario with humans and AI, (which also is likely to lead to a bad outcome), the world needs a radical breakthrough in input/output bandwidth, so humans can become "one" with AI should a person desire; hence Neuralink.

TL;DR: Superintelligent AI is coming whether we like it or not. Neuralink's mission is to minimize that existential threat. Should be an interesting ride.

 

[Intl Professor]

I forget his name, but in the very next issue of Scientific American after this piece by d'Espagnat the most famous theoretical physicist at MIT in my time angrily took it apart. If I remember correctly his main criticism was the evidence d'Espagnat used was merely due to the fact everything in the universe was in contact at one time at the time of the big bang.

In the arrogant delirium of my dotage I would say now, "sooo, professor, isn't that d'Espagnat's point? QED." We've been stuck with that intimacy ever since for better and worse. (I like "and" better than "or" out of respect for Hegel's first law of the dialectic.)

Edit: Sorry I couldn't find a link to the Sci Am article and rebuttal, but I did remember the physicist critic's name: Victor Weisskopf. In my time at MIT ('54-'58) the 'toot had no indigenous Nobel laureates and the administration was somewhat envious of Cal Tech. I learned from Wikipedia why he did not win a Nobel prize for a discovery someone later did duplicating Weisskopf's earlier unpublished work. After the link are some memorable quotations for which I did not know W. was famous. You might like them this day.

Victor Weisskopf - Wikipedia

"He was appointed by Pope Paul VI to the 70-member Pontifical Academy of Sciences in 1975, and in 1981 he led a team of four scientists sent by Pope John Paul II to talk to President Ronald Reagan about the need to prohibit the use of nuclear weapons.

In joint statement Preserving and Cherishing the Earth with other noted scientists including Carl Sagan it concluded that: The historical record makes clear that religious teaching, example, and leadership are powerfully able to influence personal conduct and commitment...Thus, there is a vital role for religion and science. [15]"

and,

“ Human existence is based upon two pillars: Compassion and knowledge. Compassion without knowledge is ineffective; knowledge without compassion is inhuman.[18] ”

"In class one day, speaking to junior physics majors (Spring, 1957): 'There is no such thing as a stupid question.'"

I remember something he said at a public lecture but have to stop now, my wife misses me.

 

[Intl Professor]

Do you remember what can happen when you're deeply in love with someone: on occasion a shared experience appears to create a simultaneous bond? Each of you feels (there's that imprecision of language - right there!) that empathic connection. That is what I have been envisioning when I read these lines.

I get flashes of that every time my lovely Buddhist wife nails me with a so-precious smile as a hypocrite. Previous loves fall under the category Maxim Gorky called "a madness due to a failure of the mind to comprehend the world." My dear sir, you are blessed in so many ways.

 

[Intl Professor]

snip
“ Human existence is based upon two pillars: Compassion and knowledge. Compassion without knowledge is ineffective; knowledge without compassion is inhuman.[18] ”

"In class one day, speaking to junior physics majors (Spring, 1957): 'There is no such thing as a stupid question.'"

I remember something he said at a public lecture but have to stop now, my wife misses me.

A famous Yale professor of philosophy came to campus sometime in the mid to late 50s. It was in a small lab lecture room, one with tiered seats so everyone could see what was going on at the table used for some experiments. I was seated in the first row within reach of the table as the distinguished professor shouted while pointing at an ashtray "what a horrendous assumption it is, that it exists." As a young engineering student I felt like picking it up and throwing it at him to see if he ducked. (Fortunately, I was too timid and did not know about operational definitions so critical to Einstein's thinking, although it sounds intuitive now.)

We all gasped and looked up to the last row at the top, in the rafters, so to speak, for a bolt from Olympus where Victor Weisskopf, our Zeus, lo, maybe even the Mother of all Minervas, for response. He said: "it is important to see all of the presentation before comment." My memory is blank on the denouement.

Still, wish over sixty years later I could heed his advice.

 

[Intl Professor]

more later- or maybe I'll just wait until we can neuralink

Don't wait, I have little time left and I'm greedy for "INPUT, INPUT..." as the robot said in some movie.

 

[Ulmo]

All I have to say about the topic is that Neuralink and the Brain's Magical Future - Wait But Why is pretty complete for now. He's very slow to develop the concepts, but it's done well even for the speed he goes, and it is done in such a way that anybody who can sit through the exercise will understand it, bringing us all up to speed. I'm even thinking of handing a copy to my mother (in actual paper book form), even though she's an anti-future anti-technology anti-intelligence anti-learning Luddite; it's that well written.

It is long. I skimmed most of the stuff I wouldn't learn anything about what I knew others had figured out, and stuck to the stuff that revealed what they knew that I didn't know they knew. It still took me many hours to make sure I caught it all. Even though I have my usual dismayal about the languid state of affairs, the piece is so complete that even that dismayal is pretty well assuaged by the many things listed that have been done. But more importantly, the concepts, reasons, and goals are pretty well fully described, and often, that's the part that is missing in any description of a topic. This one is worth saving, since from now on, everyone will probably not do the work necessary to fully describe the concepts contained therein. The next time that full description is fully made available to humanity will probably be when we do get to the hive mind state and someone saved a copy somewhere and we just read the old one again (and at that point at least can constantly update it from there).

 

[winfield100]

For a (partial so far) discussion of neural link (i am only at page 229) consider reading some reference material
Iain M Banks, {author} chaper 7 of "Consider Phlebas" 'a game of damage' for how (with consequences) of a neural link of a few 1,000's with security guards, players, aliens of many types, actual deaths, "who am i and what are my thoughts vs ...." problems caused by sensory overloads and mind overlaps, mental weapons of emotions, singular and multiple, betrayals, etc.
{wait but why is a condensation, paring down}
Iain M Banks IS where Musk got most likely initial ideas and "what if's"
{if you lived during the 1960's, think purple haze [200 mikes+], with orange barrel and blotter with on/off/intense/mild/wahhoo!/oh sugar! switches that occasionally malfunction or maybe didn't or did or..}
[you have to recursively read the book, so it is longer than it appears]

 

[Skipdd]

I have initiated this thread to be the initial catch-all for discussions involving Mr Musk's new venture, Neuralink. In short order, we hope, we'll have a permanent sub-site for Neuralink.

At the outset, we'll most strongly suggest (from Lord Vetinari's Dictionary: see "DO") that before chiming in, all participants read Tim Urban's "Waitbutwhy" blog post regarding same: Neuralink and the Brain's Magical Future - Wait But Why

My first impression from that exercise is that when considering Neuralink, I had to make a comparison with Tesla Inc. and with SpaceX.

As follows:

Aaaww, aren't they cuuute!

Thanks for starting thread and posting the tutorial. As @Ulmo said, article is very accessible and understandable even for those not having technical training in the areas it covers. Fascinating and de-mystifying read.

 

[Intl Professor]

The competition heats up. Here's another approach by Zuckerberg.

The smartphone is eventually going to die — this is Mark Zuckerberg's crazy vision for what comes next

 

[Ulmo]

This sort of thing has never sat well with me. Computers can get hacked and I really don't like the idea of people's brains getting hacked.

As the WaitButWhy posting said, that's going to happen anyway, so the idea is to position us in the best possible position to be able to handle it when it does happen.

I keep telling everyone that if you want to go into anything having to do with computer programming and/or artificial intelligence, the best shortcut to a great understanding of those things at least in the near past was to start with a full understanding of biology, including all the wonderful functions of the body, disease, etc., because most of the concepts in programming today have the same types of complex thoughts and learning associated with them. Even the US Navy (sp?) knew this decades ago with their DARPA proposals, and I was glad to see at least someone knew it back then when they came out with those proposals (I was like "finally!"). Anyway, the reason I mention any of this is: a lot of the same concepts in fighting disease in biology also exist in such a conglomerated human brain network. You have to have a lot of pieces that redundantly handle disease. If you instead take the approach that you don't like disease and therefore do not want to have the pieces that might get diseased, such as a human body, then what you end up with is -- you don't exist. I jumped from a smaller conclusion to a larger conclusion. Let me try to break it down to my initial attempt to describe it. If you are afraid to integrate the parts of the human body because that allows more diseases to infect more of the body (such as through the blood stream communicating the infections), then you are also afraid to have the systems that help fight those diseases throughout the body (such as through the blood stream communicating the battles). If you are afraid that together a total system can be attacked with leveraged consequences, then what happens when a non-system is attacked one bit at a time incrementally and none of the individuals have control over the infections of their isolated pieces? Eventually, all members of the society die. For instance, look at a conglomerated society such as USA+Europe fighting diseases with such things as vaccines and antibiotics, vs. societies that don't have that because they stay independent, and sometimes get some portion of wiped out (even up to including fully wiped out).

I'm not saying it is a guarantee to be better, but I am saying that it has more potential to be better. And, this is on a huge scale, so more equals much much more.

I do find it interesting that I hate the problems of collectivism and communism, and yet, I am fully in support of some portions of the systems taking advantage of conglomerated progress; this is fully explained in this paragraph and the next by individualized subsystems. As in most things, balance is what we need. If you take the concepts I describe in the next paragraph, bringing structure (such as (quasi-hierarchical or other structured) sets each with self-protections) into the systems is the balancing force. I think being able to have more sooner understanding of this hive is better by being in it. Otherwise, as Elon said, being a pet is a good outcome, and I'd rather, as Elon said, be part of the outcome's future beyond just being a pet.

Much as nations have as their body's skin borders, humans have as their body's skin skin. Oh, I just took an analogy that I applied to a nation from the human body and re-applied it back on the body. That's slightly funny. But my point is that this can be extended. Cells have cell walls. It goes down further and up further. Earth has earth walls (well, space). Elements have strange forces that hold them in sort of limited ways. You can go on and on, like states, counties, mountains, continents, organs, vessels, pipes, neurons, etc. Nothing is completely impervious, yet, that resistance element allows individuality, within greater conglomerated systems, so what happens is you get a bunch of individuals that are not completely violated and have diverse reactions to the system as a whole, and the system can be made better from the effect of uncorrupted (or at least not too sick) individuals and the individuals that can be made better by the uncorrupted (or at least not too sick) system (or even subsystems). Thus we live. (Probably.)

 

[Intl Professor]

Don't wait, I have little time left and I'm greedy for "INPUT, INPUT..." as the robot said in some movie.

The movie was Short Circuit.

 

[Ulmo]

If it ever comes to fruition, I pray it is never applied to my MS OS. Too often I've thought; "I should just ram that SOB who just cut me off with inches to spare..."

And I thought AP2 was trying to kill me every few minutes...

It doesn't take much discipline to create a safer more mature thought version of that and just re-analyze everything in life. As long as you remember that, and do it when it becomes necessary to not be overbounds, then you'll be fine, I think. However, I don't know: I cringe every time someone says it's ok for imprisoned persons to be tortured, killed, or raped by (other?) criminals: that thought is an example of how corruption of thought turns into corruption of action, and the bound not to cross. If people continue having those thoughts, it's as if they plugged their brain into the MS OS (Tesla Model S car software) and had those thoughts to ram other cars, but never stopped their car from actually doing the ramming part, and then rammed.

 

[ecarfan]

Just finished reading Tim's Neuralink piece. It's a lot to take in, but as usual he does a remarkable job of clearly conveying the necessary concepts. Of course some of the concepts are beyond what one can imagine because there is no real world context to relate them to. Those of us who have spent a lot of time reading high quality science fiction may have an easier time of it.

My takeaway is that while I can dimly imagine such a BMI someday being a reality, the timeframe seems far longer than the timeframe to create the super powerful General AI of the future that seems to be the reason Elon created Neuralink. Which is unnerving.

I was also very impressed with Tim's graphic explaining the process that drives Elon's various companies (Credit: Wait Buy Why)

(Credit: Wait Buy Why)

 

[SW2Fiddler]

Banks' Culture series starts from a base of 'good outcomes' from free energy (post-scarcity) wielded by benevolent AIs. That's all a given and NOT given a ton of exposition. The plots involve what the 'house cats' do in opposition to the ease-of-life, or to the rules of the game, or... well, just read 'em.

 

[SW2Fiddler]

I read the article, next I will go back and read the linked articles and watch the videos. They were threatening to sidetrack me from finishing!

Consider that homonculus. Don't underestimate the bandwidth accessible thru the nose, tongue, lips - see Proust (or the film Ratatouille) for a nice example. Good chefs are already efficiently transmitting very complex data from their brains to ours.

 

[SW2Fiddler]

...a lot of the same concepts in fighting disease in biology also exist in such a conglomerated human brain network. You have to have a lot of pieces that redundantly handle disease. If you instead take the approach that you don't like disease and therefore do not want to have the pieces that might get diseased, such as a human body, then what you end up with is -- you don't exist.

This very comment should also be Assigned Reading!

 

[ecarfan]

Consider that homonculus. Don't underestimate the bandwidth accessible thru the nose, tongue, lips

Those sense organs evolved to detect specific types of environmental inputs, but not the information used for higher cognitive thought processing as done by humans. Dogs navigate their world by smell, humans use their sense of smell to a much more limited degree.

I don't see how those sense organs would be useful as bandwidth conduits between the human neocortex and the information and processing capabilities of a general AI system. Please keep in mind that, for example, the nerves connecting the nose to the brain are relatively few in number. Neuralink is trying to develop an direct interface between a general AI and the neocortex. Very different than using human sensory nerves to the neocortex.

 

[AudubonB]

Here is another reference point all might consider - today's (Sunday 23 Apr but it appears to have been released 21 April) TED Radio Hour was...

The Digital Industrial Revolution
Released Apr 21, 2017

As machine learning surpasses human intelligence, where does that leave us? This hour, TED speakers explore ideas about the exciting and terrifying future of human-robot collaboration.


I know there are lots of ways to access this, for free; here is one: TED Radio Hour by NPR on Apple Podcasts

I missed some 30-40% of the talk as I had it broadcast in my workshop but I was struck by how many of the interviews provided gave me the same reaction: "Whoa! We're already beyond that. What you're saying and proposing is so 2016."

To summarize, then, we can take a maxim from another field (many, many fields): Time is of the essence. There is not a moment to lose.

 

[Ulmo]

As the WaitButWhy posting said, that's going to happen anyway, so the idea is to position us in the best possible position to be able to handle it when it does happen.

To beat the dead horse:

Elon Musk on Twitter

==== Copy of above link ====
Juan Carlos‏ @juancarlosrs

Is this a counter for skynet? @elpilot #robotuprising god bless you @elonmusk

Elon Musk's latest vision for the future? Linking human brains with computers @emilycanal Elon Musk's Neuralink Wants to Fight Brain Conditions Using Cloud-Based AI

7:56 AM - 23 Apr 2017 2 replies 1 retweet 7 likes

++++ Elon's response ++++
Elon Musk ‏Verified account @elonmusk Replying to @juancarlosrs @elpilot

That is the aspiration: to avoid AI becoming other.


10:23 AM - 23 Apr 2017 29 replies 25 retweets 200 likes

========

I'll repeat what Elon said Neuralink's aspiration is: "to avoid AI becoming other.", as in, "to avoid AI becoming other than ourselves." It will anyway, as its own beings, but he is talking about other than what we can coexist with. I am wrong about what he's thinking; I'd like to hear it more. But that's the gist of it in a rough way.