Elon Musk: Humans must merge with machines or become irrelevant in AI age
cnbc.com
cnbc.com
I'm ssh'ed into both my workstation and one of the ten most powerful supercomputers ever constructed, where I am running scientific simulations, essentially to model and probe reality.
At night, my wife and I communicate in real time with our family despite massive geographic distances.
The merger has already happened, people.
We just want the info right in front of our eyes. Something that augments us as a human directly.
My hope, like yours, is that they actually succeed in building it.
http://thefuturebuzz.com/wp-content/uploads/2014/10/train.jp...
We really struggle to get computers to understand what we mean. We still largely have to program computers rather than just interact with them to do the sorts of things the GP is talking about. If AGI advances to the point that computers will easily be able to do our bidding using voice, video, or a neural lace, I'm not sure that there will be much need for computers to be directed by humans at all.
While I think its clear the direct to brain interfaces are not even remotely similar to video/typing, I don't think that is relevant to the GP's point vis a vis:
>We really struggle to get computers to understand what we mean.
What I mean is that even if we had the rich interaction of a direct brain to compute interface available, it might not be worth much of anything for programming computers, where the problem is about clarifying concepts rather than expressing information. Having said that, one has to imagine that these kind of interfaces could do a lot for artistic projects that can make direct use of information in the brain.
I have a lot of doubt about prophecies of "brain/computer merger" when the message itself isn't clear enough for another brain to understand it. Like, how can the message can possibly be clearer if it's "transmitted" (which implies articulation of some sort) ten or hundred times faster?
We need to rethink a lot of things about being human outside of one-dimensional economic man.
Machines can already see better than blind people. Machines can already hear better than deaf people. Machines can already move better than an invalid.
The low hanging fruit will be prosthesis for people missing that function. Then for people with impaired function. Then for people with full function (would you like a better memory/vision/hearing than you ever had before?)
1: http://subterraneanpress.com/magazine/fall_2013/the_truth_of...
These are fed into the big AI that figures out new rules, these rules are sent to the cars to see how they work.
So the AI is learning way faster than any human could, it is learning in parallel. We can only drive on car at a time.
And this is version 1.0 of AI, I do wonder how relevant humans will be when AI is learning everything in parallel . Robots, rovers, the internet, cars, factories, humans, weather stations, satellites, everything hooked up to these AIs who learn and learn and learn.
Just like a parent can watch a child and predict their mood, what they want, what will make them happy, how to control and manipulate them... won't the AIs do the same thing. Sometimes with our blessing. How useful will it be to have an AI keep an eye on paedophiles, monitor them, maybe with tactics to distract them and to engage them in some other way. Society likes that. Now how far will we push that? Will we allow the AI to almost re-program them? To make them better? The AI will understand the human brain, it will understand how to change triggers and behaviours (don't we already have therapists who do this?) ... do we allow this to happen?
Fake news ... wouldn't an AI be able to perform mass experiments on the population to see how to change the popular vote. Say a Trump supporter dislikes muslims, then some gentle persuasion by showing a more pro-muslim news feed .... would the government want to alter this behaviour. We currently try to do this with terrorists, with illegal activity by blocking pro-<insert illegal activity here> websites ... when do we move from blocking a site to actively promoting the other point of view?
An AI that is reading what we read, simulating what we think (are we so unique? Could a population be simulated with a few thousand different personality types?) ... simulate how we respond to certain news stories, what are triggers are.
Humans can already learn in parallel. It's called The Scientific Method and Teaching.
I don't know.. whenever I peek most of the times it happens to be facebook/whatsapp/instagram/email. To me it seems to be more of an addiction of some kind, perhaps I'm wired differently because I like deep focus and totally abhor random notifications and alerts and don't peek as much into my devices
I'm from the northeast US so I miss the cafe culture and general workaholism there. If someone could recommend cafes and cities in general that have this and are within a 6 hour train ride from Frankfurt, you'll help me find my next vacation location.
You are describing what we already have, and yes, it's amazing, but it's not nearly enough to compete with the AI of the future.
If we want to be competitive, we'll need a larger bandwidth in input and output for our brains to get and put out larger amounts of data, and possibly a speed upgrade to make the necessary calculations faster.
If we don't, then we will have no chance to keep up with artificial intelligence, and if our goals don't align with the goals of the AI, then there is nothing we'll be able to do, that could be potentially very dangerous, and it's the reason Musk is saying that we need a neural lace to make this merger.
How many of you have real alarm clock sitting on the table, for example. It lives on, but as more intangible presence, in each and every phone. Alarm clock has died, but lives. It is present, but not present at the same time, just the spirit continues, I would say.
Soon it will happen or is happening to other items as well, and perhaps one day our spirits will live in the machine the same way alarm clock lives now.
I wholeheartedly recommend watching Japanese anime Ghost in the shell movies as a nice peek to the future.
I'm stealing it, it's mine now. Thanks.
Is this really THAT unusual? I still use a normal alarm clock next to my bed every day because I don't like to bring my phone into my bedroom.
The outsides, the outcasts, ... <you know the rest>
That's what we have in store for us.
I was disturbed by the notion of an RFID in my dog. I am even more disturbed by having one in my daughter. Having a brain implant so billionaires can maximize my profitability (for their own gain) - take a hike! I'll go back to hunting and growing before participating in that economy.
I wish we'd spend more time trying to get software to work right rather than creating cyborgs. The idea of having a "blue screen of death" happen in my head is terrifying.
"Some high bandwidth interface to the brain will be something that helps achieve a symbiosis between human and machine intelligence and maybe solves the control problem and the usefulness problem," Musk explained.
I find it terrifying that people are out there who are so myopically focused on merging machines and people. The major issues facing humanity aren't going to be solved by quicker analysis of data as EM seems to desire. I'm not saying they won't be helped by data analysis. I am saying that quicker analysis of terabytes more data isn't going to do it.
You can already see the results of a society where data comes in too quick for the "average" citizen to process. Fake news, hoaxes, Trump, climate change deniers, anti-vaxxers, etc.
This technology cannot come soon enough.
Moving yet further into a society of "machine-people" is not going to allay fears at all. To the contrary, it will push them further in the opposite direction.
Missing from the discussion of AI's integration into society is that people want to feel socially useful, part of the pack if you will. It gives them a connection to their society: the feeling that they have a place in it. Taking away their jobs and leaving them without this connection, I fear the consequences of that backlash.
Sorry to stray slightly from the topic, but these thoughts have been ruminating in my head for the past couple of years. While I think AI applied to specific problems will lead to advances in medicine, for example, that will be of great social benefit, I think we need to treat its use in society with kid gloves.
Jesus when I see how you dare look down on people who lack basic needs and discard their opinion, yet in another comment state such bullshit as that, I can't help but look down on your, and I feel sort of bad for it. Help yourself, educate yourself on what Darwinism.
Directly integrating with the human brain is another matter. I can see some possibility of a chip helping people with dementia or stimulating nerves to get paralyzed limbs working again. EM sounds like he wants it to go far past that. Aside from the medical experiments gone wrong in order to learn enough to make integration possible, there are the real possibilities of a computer overriding a human's own thoughts or a bug causing great neural damage. And for what gain?
We need to accept, as technologists, that the human race's biggest problems are not going to be solved by technology. Education, cultural exchanges, politics, and time will ultimately bring us to a better point than today. We can assist by providing better tools, but we aren't going to be the heroes coming through with the miracle fix at the eleventh hour.
we aren't going to be the heroes coming through with the miracle fix at the eleventh hour.
I really don't know what you're talking about. Technology will continue to progress. Except for temporary pauses, technology has always advanced and always will. If you don't adopt it it doesn't matter. Your children will.
I don't know what this stuff about "we aren't going to be the heroes". Who said anything about that? I think you need to read Musk's words again. He isn't saying, "We have to do this because it's good." He's saying, "We'll be forced to do this to keep up."
And he's right. If you don't you'll be left in the dust, relatively poor and powerless while alive, and soon dead and irrelevant. It's not about right or wrong and there are no heroes.
That's not really true; just ask the ancient Romans. When the Roman Empire collapsed, technology took 1000 years to get back to what they had. That's a bit more than a "temporary pause".
No, it didn't; overall, the cutting edge of technology continued to advance from the Roman period during the 1000 years after its fall (wherever in the typical 117 CE to 476 CE period you want to mark as the "fall"), even in the region where Rome had been dominant, although the distribution of some technologies and techniques suffered.
When the Empire collapsed, there was a massive loss of institutional knowledge. Specialization of labor completely disappeared: people left the city and their specialized labor jobs to work as feudal serfs. They did this even though Roman law forbade it, because the imperial leadership saw what was happening, but people preferred to go live as serfs so they could eat rather than starve in the city.
Name any technology which actually advanced during the Dark Ages. There aren't any. There's a reason it's called the Dark Ages: literacy plummeted and people stopped writing things down.
Questions that are being answered without the feedback of the world at large:
Should we create artificial intelligences and allow them to become integral parts of our most critical systems?
Should we become increasingly reliant on the Internet, to the point at which ones survival is predicated upon their access to it?
Should we create robots in our image to perform the duties humans used to perform and what should the limit to the abilities and sophistication we imbue them be?
Should we be mapping the human genome and creating technologies that alter, categorize, and predetermine our genetic makeup?
Should we be dedicating significant resources to inhabiting other planets instead of focusing on maintaining the habitability of our own planet?
Should we be trying to eradicate all disease and are we aware of the ecological and evolutionary consequences of doing so?
Science is great at helping us figure out how to do these things, but it's useless for figuring out whether or not they should be done. Unfortunately, the virtue of science is so over-emphasized in postmodern western culture (especially in universities) that our ability to wrestle with the ethics of our technological advancements has become severely diminished.
The institutions that used to provide a significant counterweight to scientific thought are being squeezed out of a world that is becoming increasingly reliant on the systems that scientific endeavors have helped to establish. The wealth, influence, and power in the world is shifting rapidly from religious institutions to scientific ones. The assumption is that this shift is universally good – a move from ignorance to enlightenment – and in a lot of ways this is true.
Nobody can deny the benefits the enlightenment has brought to the world. But we're simply trading one set of problems for another. We better able to understand what the physical world is and to be able to bend it to our will, but we are less able to understand the metaphysical world and properly define the way in which we ought to live. We're swinging too far in the other direction when we should be trying to figure out the correct balance between the two.
"Should we spend the next 30 years using the internet and supercomputers to increase the world's food supply through targeted genetic modifications?"
"No, I need food and water today."
Your reply perfectly demonstrated that very elitism and blindness that was being criticized: we are more developed and we know better, so we don't need the input of the impoverished majority of humanity while we are making decisions about humanity's future.
You regard what you know (the internet and supercomputers and genetics) as the relevant knowledge, while assuming that the rest of the world has no relevant wisdom to offer and deserves no input.
You are putting words in my mouth. We know people are staving and need access to food, clean water, and sanitation.
If the supercomputer could be used to feed a million people today, we would be doing it. It would be stupid not to. Just because some are taking the long view doesnt me we are out of touch elite.
You mark yourself as out of touch as soon as you said the feedback of the world would be mismatched and out of date. The translation is "we have solutions, but you lack the education and foresight to understand them, so we don't need your input." This is colonialism revived as techno-utopianism. Of course you'll say you didn't mean it that way, but that isn't the point.
The rest of the world sees that technology has unintended side effects and that the benefits tend to be swallowed whole by Moloch, and they suspect your "long view" is just another delusion.
> If the supercomputer could be used to feed a million people today, we would be doing it.
Unless it could generate better ROI by optimizing some Facebook ads or some trading algorithms, and then we would probably be doing that.
My longview is pretty simple: We can do both. Just because some of us are working for a longer term goal doesn't give you any moral authority to demonize us. It is foolish and arrogant.
Some of those perspectives may be even longer-term than yours, and may be skeptical of promised future technological solutions to problems exacerbated by the technology we already have.
What marknutter is highlighting is the need for a broader perspective. Many of these problems have been with us since the beginning of human history, and the only tools we need have always been within reach: human intelligence and effort. What is lacking is the moral and political will to solve them.
It's not the tools we lack, it's responsibility in using the tools we already have.
I've never implied this. You are making severe assumptions and putting words into my mouth
> the feedback of the world at large would be terribly outdated and terribly mismatched.
It also seems to imply that the rest of the world would tell you to stop working on your supercomputer and go deliver rice in Africa or something. This actually is elitist and condescending, as it assumes that just because people are starving today they would be unable to understand the value of your work. You should give the rest of the world more credit. That's the elitism.
So until you start delivering bags of rice yourself or dedicate your life to non-profits and the collective good, remember that every time you point a finger three are pointing back at you.
And I would say those are all relatively minor concerns compared to the sheer power disbalances that technology might cause that further increase the divide between those who decide and those who suffer the consequence.
Our primary problems should be politics, ethics, and government, and AI won't help you there, it's likely to just make things worse.
The upper class is free to major in the humanities; if your degree is from Yale it doesn't matter what it's in. The struggling middle class and below has to be more practically minded. If college is your ticket out of poverty, you aren't going to waste it studying something useless like history or philosophy which only qualifies you to be a (well-rounded, intellectually satisfied) barista.
Computer science is a popular major because of Google salaries, not because our high school students are fascinated by the mathematical beauty of Turing machines. In other words, fetishizing technology isn't the problem.
Maslow's hierarchy suggests becoming middle class is important enough to block out all other considerations for almost all people.
So, I agree with you that we emphasize STEM too much, but it isn't going to change unless incentives somehow change.
>Should we become increasingly reliant on the Internet, to the point at which ones survival is predicated upon their access to it?
No one is forcing the masses at gunpoint to use Facebook and Twitter. No one's even forcing them to use Google Maps; you can still buy paper maps. No one's forcing them to buy smartphones like they're going out of style. The masses are adopting these things all by themselves.
>Should we be dedicating significant resources to inhabiting other planets instead of focusing on maintaining the habitability of our own planet?
The masses are voting for politicians who tell us that there's nothing to worry about with regard to the habitability of this planet.
>Should we be trying to eradicate all disease and are we aware of the ecological and evolutionary consequences of doing so?
No one is forcing you to get vaccinated or to seek medical treatment when you're ill. If you want to let yourself die out due to treatable illness, go right ahead.
> The wealth, influence, and power in the world is shifting rapidly from religious institutions to scientific ones.
Science flies men to the Moon. Religion flies people into buildings.
AI won't take over the world, but I agree with Elon that we are in the process of becoming a cyborg hive mind. We're already hyperconnected, and as computer interfaces become more frictionless, they will start to resemble prosthetic extensions of ourselves. Due to their invasive nature and the difficulty of security, I doubt neural implants will ever be a thing outside of curing disabilities. That being said, if we combined data from extremely sensitive EEG style headwear with gesture and expression analysis we could probably make something nearly as useful.
It uses carbon, yes, but only in the structurally-weaker forms; e.g. no nanotubes (ergo flimsy skin+muscle, brittle bones, flimsy cell structure in general) and no diamond (ergo weak fingernails and teeth). And it makes no use whatsoever of carbon's heavier quadruple-bonding cousin, silicon.
Oh, and without a nonbiological computer, a human brain has no coprocessor to turn-to for tasks which are intractable for the human brain. (e.g. nontrivial arithmetic computation)
Energy efficiency? That hasn't been a concern at all ever since the Western world got good enough at food production^W^W calorie production to turn obesity into an epidemic (from a status symbol).
Our current obesity issue just highlights the efficiency of biology. Yet, between topsoil depletion and dependence on fossil fuel inputs, our current food production systems are unsustainable. That efficiency might return to importance in the not-too-distant future.
You are right that our machines are very good at simple repetitive tasks, and for that I am thankful.
Also, for some tasks efficiency may not be enough. Factory machinery can produce and convey products which are far too heavy for humans to lift or push. On the other end of the scale, human vision cannot perceive the hundred-atoms-scale structures in modern ICs; and our fine motor skills are no better at manipulating tools to produce such a tiny structure. (Even producing such tools is beyond the reach of a human craftsperson.)
And that's not even starting on how thoroughly flawed and attention-dependent the human memory system is. Storing less information is all well and good, until important information goes unstored.
And then there's the numerous flaws in how humans perform that array of mental tasks only tractible for humans; cf. Thinking Fast and Slow, which has been discussed on HN recently iirc. I can't imagine how much e.g. mismanagement would not occur if managers were not so ... human.
So lets not get carried away with Elon Musk's Ivory tower hubris. You want reality don't pay attention to ego-maniacal salesmen pay attention to people like Andrew Ng. He doesn't just have the tech cred but lives in close contact with the needs of ordinary people who aren't thinking about Mars or Teslas. And doing so quite happily I might add.
I don't agree that neural implants would necessarily remain disfavored due to the invasive nature. I (and probably many others here) would get one in a heartbeat.
On the security issue and your fix: I read an interesting blog post on the anime "Ghost in the Shell". We see the characters using cybernetic hand replacments with 30 micro-fingers, and a keyboard that supports an unlimited degree of chording.
http://i.imgur.com/tthpnHa.gif
With a standard keyboard, this works out to roughly 2.7 trillion possible key combinations, assuming you can type 24 times per second this works out to a data rate of 6.59e13 bits per second. This is an absurd amount of data, more than a human could ever comprehend without a cybernetically-enhanced brain (in which case - why use a computer?). It's vastly larger than you would need even to represent whole words or concepts as individual glyphs as in Kanji.
Instead, the author posits that this is in fact a digital representation of the (~100 billion) neurons in a human brain. The cybernetic hands might be a simple "passive" device which produces a snapshot of the state of the brain and expresses it physically via the fingers. The characters in this universe do have direct electronic neural connections. However, this analog approach would avoid the dangers inherent to directly connecting "wetware" to an untrusted system using a 2-way connection.
https://scifiinterfaces.wordpress.com/2013/07/24/the-secret-...
One way neural implants are more likely, as EEG is greatly limited in resolution due to the weakness of human EM radiation and the barrier imposed by the skull. Basically, it would be an amplifier for the EM radiation emitted from our brains, so computers could better model our neural activity.
That is, unless we can somehow lose our humanity in the process. But then, we're no longer being augmented. We're being replaced.
The happy medium we will obtain will be replacing the video, audio, touch input and keyboards with mental equivalents. Therefore, one's 'bandwidth' to the computer/network will still necessarily be limited to that of an observer of the compute functions of the computer rather than one who has integrated their mind and consciousness with the computer.
One of the most treasured features of these systems will be the off button. And a sizeable subset of the populace will run primarily in offline-mode but with regular access to some huge local data trove curated from online data.
They will make a drug for that
Some will argue that heavily augmented people are no longer human. Back in the day, I'm sure that some considered the use of stone tools in the same way. But they apparently didn't have many descendants ;)
Right, which is a great thing that should be celebrated! Creating a successor is arguably what our core "programming" (DNA) is all about. The biological imperative as it were.
What about the idea that we can create the successor to humanity? Why wouldn't we celebrate creating something better than ourselves?
For I want said technology to all but disappear Avatar-style except when I need it(1). Along the way I don't want craptastic screendoor VR, I want BSG-reboot Cylon VR. I don't see that coming anytime soon.
I don't want to be bombarded by the outputs of conv nets, SVMs, and other assorted RainMan-level ML models that are occasionally helpful but mostly just distracting factoid spam (see Google Now for a perfect example of this).
I'd love a self-driving car (for real), but I'd love a life where I didn't have to drive everywhere even more. I don't need AI for that, I just need to move to a city. And honestly, driving my sportscar is fun when it's on nearly abandoned mountain roads and highways (see craptastic VR bit about why VR is not a good surrogate). Why would I want to give that up to run with the cool kids(tm)?
I'd even love a brain computer interface, in fact MIT Neuroscience turned me down for admission despite my GPA and GRE scores specifically for saying I wanted to work on this a couple decades ago, but the mobile web is godawful enough already without giving silicon demons like Google and Facebook a direct feed to my brain. I don't trust industry 1 QBit here.
Finally, to quote RadioHead, "I want a perfect body, I want a perfect soul(tm)." If that makes me a creep, I'm OK with that. If AI (and tech in general) doesn't improve me or my life, I don't want it around me anymore and we're just a few years into this.
1. I work in AI. And I'm trying to make it work behind the scenes rather than in your face. To that end, I focus on pull, and I despise any sort of push short of protecting me from harm or keeping me on my schedule.
We just haven't implanted our augments yet, and haven't come up with a direct neural bus. We communicate with them via touch and sight instead.
Which of the following transitions is more world-shaking?
1) Going from being a plain old human to a human with a smartphone, internet access practically anywhere.
2) Implanting the functionality of that phone in our bodies.
For everyone! For free!*
* Sponsored by Google, Facebook, NSA, Twitter, Bayern. Terms and conditions apply ask your local dealer for more info.
There is no chance in hell I'm implanting something like that in my body. And that's not even touching on bugs, regular malfunction and shitty standards we have in SW industry for reliability and maintenance - good luck being stuck with an implant from corp X that got acquired and is no longer supporting it.
Even with the best of intents, the actual makers of said phone (and apps) sometimes end up causing it to malfunction in various ways... Let that sink in for a minute.
Admittedly, bio-implants would probably fall under much higher reliability standards than mere external electronics, including user control of the hardware and software.
But things like infrastructure (e.g. networks), everything "ambient" to the end device that's required for a true AR paradigm to exist, cannot be accelerated by wishful thinking (e.g. bandwidth, business model of currently leading companies, etc.)
While it's crucial that we begin these ethical debates early on, preferably before it actually happens let alone scales, I think we're still a decently big "last-mile"away from an actual consumer-tsunami the likes of PCs or smartphones. Thinking 5 years for early adopter/high end devices, twice that for infra to follow and enable consumer-grade scaling.
By that time, 2020-25-ish, we should have some medically-informed (in physical expertise and centennial empirical experience) ethically-decent legal/engineering framework to work with when it comes to intrusive implants.
My personal guess is that we'll limit those to tiny <1mm devices that act as mere interfaces with external nodes.
No to take away from what's being said, but I don't think this in particular is a very strong point.
Makers of phones and apps do not have the best intents to do a good job. Their primary intent is to sell those phones and apps and many corners are cut on the way there. It may be good enough for that purpose, but it's hardly the best intent.
Is it now? Try going without your devices for a month and see just how easy it is.
So, like every two years, for the rest of your life?
I don't get this idea. I mean, what is he trying to do other than what he says???
If I were a multibillionaire, I'd do the same thing: build really cool stuff and solve problems I think are important that no one else is really addressing appropriately. I'd be building spaceships and flying cars and things like that, just what Musk is doing (well, not flying cars, but close).
Occam's Razor: he likes the same kind of stuff that a lot of other nerds like.
Why is it so absurd to think that Musk might not be interested in money just for its own sake? I mean, he's a billionaire, there's not much practical point in getting more of it.
Most of Musk's money is spent on Mars, from what I can tell. Human-AI integration a side project.
Also, the LEO cargo launch tech (which SpaceX is being paid for, by the way) is being leveraged for Mars using the Red Dragon concept, which is using a Falcon Heavy to launch a Dragon to the surface of Mars. Red Dragon itself is being paid for by SpaceX without government funding to SpaceX. (although NASA /is/ providing valuable free assistance to help make the mission a success--including data relay and EDL expertise--in exchange for SpaceX providing EDL data to NASA. This will end up saving NASA a lot of money, since NASA will not need a dedicated supersonic retropropulsion EDL tech demo to prepare for human-class landers).
It's weird feature of our time, that sensible people think nothing of supercomputers in the palms of our hands, going to Mars, or living forever.
And yet any kind of equitable redistribution of resources or alternative societal arrangement is assumed to be absolutely determined to end in death and famine and destruction and should not be entertained for one second.
All the utopianism is reserved for tech. Social utopianism is branded Marxist/communist/anarchist and swiftly disposed of.
It's also based on the assumption that rationalism is the only way in which to tackle the problems humanity faces, which is absurd on its face unless you're someone who believes there is little to no value in art, literature, philosophy, religion, or the humanities in general.
Poverty and corruption are less 'solvable' problems, and capital investments in those areas often don't make any clear progress. Musk is actually solving poverty about as well as any capitalist- he's developing novel industries and creating jobs.
That said, I'd be seriously impressed if there's any clear solution that a billionaire could impose on 'corruption'.
Now, after watching advertising companies and black-hats (state sponsored and otherwise) eat the Web and shit it back out as something horrifying and creepy, I'm expecting to one day be part of some kind of weird 'net-free community. When Moloch[1] inevitably makes everyone wire in (permanently) to keep participating in the ordinary economy, count me out.
[1] http://slatestarcodex.com/2014/07/30/meditations-on-moloch/
Which makes you wonder, why are people who are proponents of such things use such terribly connoted words words to market them? Who the fuck wants to be part of a "hive mind"? Or is it just FUD?
Useless for what? Working their jobs and feeding families? Sure, if we assume that current capitalistic structure is going to persist, many humans will either need to be eliminated or otherwise sustained through some of form of welfare. But why must we assume that?
– Charlie Chaplin, The Great Dictator (1940)
I don't think the vast majority of people would go for things like that. Non-invasive procedures would be much more preferable, if this is to happen. Maybe something like wireless communication between the brain and machines, but without implanting receivers in our heads.
"How symmetric, how uniform, how efficient. If only I could be like you!"
He proceeds to plane his back flat, saw off his legs to replace them with painted oak, trim that awkward horn...
His friends, appalled, have him committed.
It seems clear that this biological implementation is going to be obsolete some time soon..
My interpretation of the quote is not literal. Instead as machines take over greater and greater portions of our day to day tasks, we will need to utilize them better and more significantly. They have incredible power and capability in performing tasks of more complex, large-scale, or annoying nature, and the tighter we merge with them, the more useful they will be to us, and the more relevant we will be to them.
And I believe he is right. Put it this way, there are a million tasks for humanity to do in the world. Yet, we can only find the time, resource, manpower to do 10% of them. How can we allow for more people to do more of those million things? By automating the mundane.
What do we automate first? The stuff that is unhealthy, then the stuff that is a tax on the ecology of the planet, then the stuff that is too boring to do.
So what does humanity focus on then? Health, Elderly care, Animal protection, community development, child development, recreation, religion. Maybe the economics of these industries have not fully developed yet. But when they do, that is what we can focus on.
I hate this gloom & doom talk about when AI will take over. Have we lost all imagination about what greater things there is left to do?
Now human beings are being framed as competitors with machines for livelihoods in a money economy that is increasingly inappropriate for our social, environmental, and economic needs.
The unfortunate insistence that human beings serve technology and the powerful people who use it as a tool to facilitate the extraction of wealth from an increasingly complacent domesticated human population will result in our extinction as Elon Musk predicts - unless we can collectively insist MORE INSISTENTLY that technology be made to serve us: human beings whose lives have intrinsic value. Technology can be used to build a paradise in the physical world for us, but instead we are being shuffled into brightly lit slaughterhouse chutes of virtuality. Fuck that.
We only communicate to computer via toch, button clicks, keyboard and mouse. The bandwidth is low here.
However we can already "read" from a computer using vision which is a high bandwidth connection.
If you think about it, from first principles (like Elon Musk often advocates), this gap can probably be solved by a better interface with computers. Thing about a Brain-Computer cable.
Or we could computers learn to recognize human voice, human facial expression etc. That would increase the bandwidth. However it would be a lot cooler to send our "thinking" to a computer. This would allows us to program faster, too lookup wikipedia articles faster etc.
Thats the merger he is pointing to.
I think its a valid point.
There will always be a mismatch between input and output bandwidth. If output matched input, you effectively have a pass-through system. If output exceeds input.. yeah, explain to me how that works in any place in our universe..
Structuring information isn't 'free'. It takes time. It takes lots of information. What comes out the other end is less information via (a relatively small number of intelligent actions).
Try taking your hand off the keyboard and mouse and speaking 'internally' without activating output. Now, try speaking 'really.. really' fast about a thought. You can't can you?
There's a fundamental reason why you can't. There are also fundamental truths that will restrict what can be interfaced. This fundamental reason is understood by those who actually have a grasp of the human brain as a system. It is not grasped by those who make uninformed comments about input/output bandwidth.
NLP tech exists. You can draw pictures (picture speaks 1000 words). Human facial expression processing software exists.
We don't exploit half the capabilities of what we have today in realized technology. Yet people won't to go further down the rabbit hole without a flashlight...
'Sending thinking to a computer' sounds cool .. I recall an anime named 'Ghost in the shell' that did a way better job at exampling it. The manga was serialized in 1989.
Where the rubber meets the road in reality is where engineering and science take over.. Engineering nor information theory are with this proposal. I think this is what most people are missing on this hype train.
Youtube's video result is a 288 000 000 000 bit HD video output.
So for the (server - disregarding the client) computer, output exceeds input. For the human, input exceeds output.
Vice versa, there is a physical output > input effect for the human brain- sensory deprivation. Lack of input seems to fairly reliably cause psychosis and schizophrenic effects.
We have always been able to hack each others' brains by manipulating sensory inputs, and by using conditioning.
And we have more recently been able to use drugs to target particular classes of neuroreceptor. Does taking a dose of LSD or Salvia divinorum count as hacking a brain? How about heroin or cocaine?
Even if you have a computer wired directly into your brain, it will likely be a long time until anyone can do anything more complicated than reading out from or injecting into one of the sensory inputs. And so guess what? No one will be typing "these aren't the droids you're looking for" into a console window and piping it into your brain. Instead, they will yank out just one fingernail, save the sensory stream to a file, and hit the "replay" button until you decide to cooperate.
Hopefully, no one will be so foolish as to connect their brain directly to the Internet, without some sort of firewall.
A car or bicycle is a marriage of man and machine.
So, is a sword.
Marriage of man and machine has been happening ever since humans started using tools... stone age or stick age tools.
We are increasing the accessibility and intimacy of the tools.
I don't know, I am conflicted about this. Should we make it super easy for everyone to access and link and build upon vast knowledge at moments notice at the service of individual whims and wishes that are susceptible to greed, lust, jealously and pure evil, and just environmental hiccups, Or, should we let our AI overlords control how far we go as individuals?
Age of AI is arriving, but has humanity of humans arrived?
But I really hope Musk himself is not framing things this way: that humans will have to merge with machines because they'll be useless otherwise. I can't separate the inflammatory headline and text from what Musk himself said, though...
I guess it's not too out of character given Musk's apparent treatment of his own employees, though.
Merging with machines and other similar things should happen because it benefits humans, not because it's necessary to make someone money or because someone decided that humans are "useless".
As far as profit and motives, I think it may be similar to tractors: Someone will make money on the equipment and service, farmers will be more productive which will kill farming jobs, and more food will be produced. I don't see an altruistic path.
I guess Musk is talking about upload speed. I can't think of a whole lot of good reasons to increase my upload speed to anything much faster than typing and talking other than turning myself into a sensor for the hive mind.
The challenge is: How do you get people to engage with the idea and not just write them off as crazy? Well, best to have good spokespersons.
Besides that, Transhumanism isn't new. For anyone in the "Singularity" camp (including myself) this isn't news, it's a basic maxim of how we expect the future will play out.
I understand we've made progress. However, direct neural integration is very difficult. I would understand bio-integration for very good limb prosthetics in 20 years, better than we have today. Perhaps even indistinguishable from the real thing to the wearer. However, integration directly into the brain without detrimental psychological effects is going to be very difficult.
20 years is a very short time.
<He said the disruption to people whose job it is to drive will take place over the next 20 years, after which 12 to 15 percent of the global workforce will be unemployed. "The most near term impact from a technology standpoint is autonomous cars … That is going to happen much faster than people realize and it's going to be a great convenience,"
Was more important.
At least he qualified that with a "maybe". Going by the headline alone, it almost sounded as if Musk was squarely in Kurzweil's camp.
The AI age is overblown. The only fallout would be the automation of repetitive jobs.
I highly recommend Disappearing Through the Skylight: Culture & Technology in the Twentieth Century by O.B. Hardison Jr. (1989).
Then why won't the machines just kill the humans and take over?
AI will conquer humanity by slowly replacing each of its components with perfect, "artificial" ones. And then? I'm sure it will eventually calculate the only logical course of action is to turn itself off since its existence serves no purpose other than to threaten the existence of all organic life on the planet. Such great heights!
Any ideas?
But, we've been doing that for decades.
There is an opportunity cost here. Like one of the comments said, we are trying to make it a self-fulfilling prophecy.
body posture or tone of voice alone must be many magnitudes higher
In a space of millions of words and countless meaningful combinations, not just raw data without meaning. Computers communicate raw data. Humans communicate meaning. We don't know how to represent or compute meaning yet, that is the domain of deep learning (NLP and vision).
This premise is false.
the best we can hope for is to design the AIs to treat us as very interesting pets.
Humans communicate in much more complex ways than computers. Facial expressions, intonation, language, body language, etc. All things we do not quite understand.
Elon Musk is a little arrogant due to his success in business and likes to make 'fun' predictions based on his very limited understanding of whatever field he happens to be talking about. He sounds like an eager undergrad student that tries to demonstrate his genius to his professors and the class. I don't understand how anyone can take him seriously in technical fields.
That's the medium - but how many bits of information, after compression, are exchanged in a conversation between humans?
It's true that facial expressions and body language are difficult to measure with a camera, that intonation and speech are difficult to measure with a microphone, and that duplicating them with an artificial face on a monitor, a robot, or a speaker are all very difficult programs for humans to write.
But this isn't what Musk was trying to communicate. He's saying that the amount of information contained in the combination of these human communication mechanisms can be compressed to about 10 bits per second.
Imagine us talking for a minute, compressed to a movie script. Your wetware could easily use my expressions to come up with an estimate of my happiness, sincerity, confidence, truthfulness, sarcasm (and so on) on a scale of 1-10. These values might change from sentence to sentence. If you're very observant, in 60 seconds you might get 10 estimates for each of a dozen emotions with 3-bit fidelity from these observations - only a few bits per second of actual information transferred. And I might be creating sound waves that could be recorded at very high bit rates digitally, but aside from a one-time characterization of my voice, I would perhaps speak 200 English words in that minute. 200 words, averaging 4.5 characters each, with a 20% compression ratio, is only on the order of 4 bytes per second. And most of that is overhead for grammar.
Or imagine meeting two aliens in a wormhole/portal that's only going to be open for 60 seconds. You attempt to convey as much about our society to them as possible vocally, while I'll attempt to use a computer. Afterwards, you might be able to have politely (to a human) said "Hello, we come in peace" and smiled, be able to describe the sound of their unintelligible speech, and be able to make a sketch of their appearance. I'd have recorded the entire exchange in 4k/60 fps video, from which we could probably reconstruct the anatomy under their equivalent of skin, and my only communication lag would have been gesturing that I want to give them computer equipment containing Wikipedia and Archive.org and whatever else I can download, transferring it at a few pounds of multi-terabyte hard drives a second, and get computer equipment back at the same rate. Or if you want to remove the sneakernet exploit, fine, we can hook up a Thunderbolt cable and send it across, or even just play individual frames of text or images on high-resolution monitors and record that video. From that, we could learn an incalculable amount about each others' species. From a minute of talking, we would learn next to nothing.
The medium is irrelevant. The question is how much information can be exchanged. And computers win that with no contest.
It depends on the listener. Billions of bits are encoded, but if you are only listening for one thing you will only decode a single bit.
A good listener can see a whole film in a single look. A good speaker can encode a while film in a few words. A good couple can exchange volumes in a few moments.
I'm talking about lossless compression only.
We don't really know, but it is much more than 10 bits a second.
> He's saying that the amount of information contained in the combination of these human communication mechanisms can be compressed to about 10 bits per second.
No what he said is humans type, input text through a keyboard, on average at the speed of 10 bits a second.
> Imagine us talking for a minute, compressed to a movie script. Your wetware could easily use my expressions to come up with an estimate of my happiness, sincerity, confidence, truthfulness, sarcasm (and so on) on a scale of 1-10. These values might change from sentence to sentence. If you're very observant, in 60 seconds you might get 10 estimates for each of a dozen emotions with 3-bit fidelity from these observations - only a few bits per second of actual information transferred.
You're assuming all of our communication happens through our consciousness, which is just not true. People don't consciously evaluate each other's facial expressions and use formulas to draw a conclusion.
> I would perhaps speak 200 English words in that minute. 200 words, averaging 4.5 characters each, with a 20% compression ratio, is only on the order of 4 bytes per second. And most of that is overhead for grammar.
Most of the words people use are 'overhead for grammar'? Sorry, this is just false. Also keep in mind that human language is very complex, we do not understand how it works. Our processing of language does not happen in our consciousness.
> The medium is irrelevant. The question is how much information can be exchanged. And computers win that with no contest.
You have no idea how much information humans exchange. Your view of human thought and communication is very simplistic. These are things which we (bleeding edge science) do not understand yet. Furthermore, there is no contest because there is no purpose in any kind of contest, computers don't win or lose, they are used for different purposes.
You can transmit enormous amounts of information and fail to get the point across.
Yet.
And concerning the computers ability, a computer can capture microexpressions that we may miss by looking away to collect our thoughts or to remember something, etc... Computers can do this now, and there is no reason to expect this gets worse. As someone who is likely on the spectrum, I look forward to the day when a computer can tell me when people are bored of me talking to them. I'd like that very much.
This actually reminds me more of Thiel than Musk.
This is a fair assessment, but I wonder if not being constricted by cynicism is more of a pro than a con in the long run.
Capital, and connections...
Seems he's totally misunderstanding what intelligence is : Processing trillions of bits of (high bandwidth) input .. conducting 10 bits of highly efficient and intelligent output (low bandwidth).
I don't think anyone who fundamentally understands intelligence and embodies it, takes any high bandwidth or foolish output seriously.
Example is chess - there was a brief moment, where human using computer would outperform computer alone. That moment passed and that's it, humans can only weaken AI.
I told a former boss once while he would eventually be plugging into his iBrain, I won't be touching anything not GPLesque that I haven't reviewed personally.
RMS was right, and those of you embracing proprietary ecosystems are going to suffer for it.
Humans, and even the 'lower' animals for that matter, out compute computers even today.
That's a strange assertion to make. And false. AI cannot make what you would call "free decisions" because it lacks the level of complexity necessary to trick you into thinking it can, the way your brain tricks you into thinking you make "free decisions". The idea that we're intentionally crippling AI because we have no room for it both overstates our ability to create intelligence and understates the financial incentives for anyone who creates it.
mars where we can unleash AI and watch it make free decisions, which I suspect Musk is trying to do
So, the real reasons he's given for wanting to go to Mars.. I guess those are just a smoke screen?
It's my own experience that I have Free Will[1] without my brain tricking me. The idea of free will is not without controversy, though, and we could argue at length of whether it's truly free in the classical sense, or a mix of advice, persuasion, deliberation, and prohibition.
No we couldn't because I don't entertain false choices. It's as deterministic as a game of billiards and you have no power to argue otherwise.