Dr. Dale Layman and ‘Robowatch’
I just ran across the humorous RoboWatch.org, operated by the nutty Dr. Dale Layman, AS, BS, MS, EdS, PhD, MOIF, FABI, DG, DDG, LPIBA, IOM. Dr. Layman apparently teaches biology at a junior college in Joliet, Illinois. He has a message to tell us about up-and-coming AI and robotics technologies. The front page explains the raison d'être:
"The reason why we are keeping a watchful eye, is because hardly anyone else outside the fields of robotics and computing, seems to be doing this! We strongly feel the need to create an organized, centrally-located, and ever-continuing watchful presence for the protection and well-being of Mankind!"
Seems innocent enough, but if you look a bit deeper, like into Dr. Layman's bio page, the "batshit crazy" aspect of the site comes into full focus. It's worth clicking on and reading the proposal and newsletter pages, with their attendant frantic and bizzare tone. Basically the idea is that Dr. Layman is afraid that humanity will become extinct due to the encroachment of AI technology. This idea is nothing new - in fact, I discuss it here pretty much non-stop. What is bothersome is Dr. Layman's old-fashioned view of AI and robotics technology, that integration between man and machine is necessarily dehumanizing, and his psychotic way of getting across the message. Obviously he is not aware of the more delicate and precise manufacturing technologies, which will one day could be used to create cybernetic organisms with as much subtlety, complexity, and elegance as the most impressive of biological creatures. I agree with him that limitations must be placed on the encroachment of AI and robotics into the human organism, but don't think that his particular approach to the problem is very productive. For a taste of what I mean, see this recent interview he did in June 2005 with Maxim magazine, available on the WTA website.
~~~
Dr. Dale Layman runs “human rights†Web site robowatch.org. And he could be the only thing standing between us and the reign of the Terminators.
MAXIM: Dr. Layman, what dangers are we facing?
Dr. Dale Layman: There are two major forces. One is penetration of the natural body, with the replacement of artificial organs and computer chips in our brains, until we will no longer be a biologically human species. We’ll become a cyborgian species. The second threat is external, from artificial intelligence.
M: How are the robots a threat?
DL: Do you think they’re going to want to be our slaves forever? They’re going to have their own rights. How are we supposed to interact with a machine that’s walking around our house carrying Grandma around? One of these machines is going to go berserk and kill someone. We’re laying the groundwork for our own extinction!
M: Are there people conspiring to bring this fate upon us right now?
DL: One is the human cyborg, Kevin Warwick, professor of cybernetics at the University of Reading, England. He implanted a computer chip in his forearm. He’s part of a transhumanist movement. The transhumanist goal is to evolve into cyborgs. He basically said, “The cyborgs will be running the show in the future. If you want to be a chimpazee, remain a human being.â€
M: How do you intend to fight the robots?
DL: I’m here to educate. I made a presentation in Cambridge, England several years ago. I dressed up as Darth Vader. I showed them the face of evil.
M: So you see this as a struggle between good and evil?
DL: This is the great mother of all battles, as predicted by the Bible. It’s going to be the fight between man and machine, between humanoid robots and their cyborgs, and human soldiers. That’s what it’s going to come down to.
~~~
That's what it's going to come down to, folks. Cyborg soldiers up in here. As promised by the Bible. The funny thing about this interview is it seems like Maxim magazine is making fun of him and he doesn't even detect it.
The Danger This Time Isn’t Human
We are used to human threats. Ignorance, mob rule, militarism, religious and ethnic divisions, etc. We hear about these civilization-wide sources of conflict from the time we start to understand language at all. They are human-style threats from an all-human world.
The all-human world is about to change. We're going to be introduced to new intelligent beings of our own design. The first will probably not be very anthropomorphic at all, because it's far too much work (not to mention unwise) to replicate the entire human complement of conflicting motivations, desires, and instincts in Artificial Intelligence. Researchers are going to focus on the intelligence part first, and skip the unnecessary baggage.
The result will be something profoundly foreign to our human-calibrated minds. Our habit of seeing the world in human terms, of course, is due to evolutionary psychology. If we grew up on a world with two intelligent species, then our brains would have evolved instincts to model individuals from both groups, and make generalizations between them. But on Earth, there is only one. Our brain has a tendency to make a category errors when contemplating nonhuman intelligences, pigeonholing them in human terms, a cognitive crutch.
To me, the concept of the Singularity has little to do with technology directly. It is a matter of creating higher intelligence, intelligence not limited by biological constraints. Technology is the means only because magic won't work.
Leading up the Singularity, there will be various risks from human sources, mainly having to deal with what happens when animals adapted to hunter-gathering on the African plains get introduced to factories that can output their own mass in product in a few hours. (Molecular manufacturing.) On the other hand, there is a significant chance that true Artificial Intelligence will actually be developed before the first nanofactories, making the risk from nanofactories moot. If so, it would make sense to focus all efforts on AI at the exclusion of anything else, because that would be the primary global risk in our immediate future. Problem is, we don't know if molecular manufacturing or AI will come first in advance, so it makes sense to focus on both.
When the first transhuman intelligence is created, the spotlight immediately shifts away from humanity to this new being. It's prudent to think about it in terms of only one being at first, because the expense and work necessary to produce a novel form of intelligence will be so great that parallel versions are unlikely to be introduced from the start. When we begin our analysis at a hypothetical point where numerous transhuman intelligences have already been introduced, we skip over what happens between the creation of the first and subsequent versions, and innocent-seeming assumptions can slip in unquestioned.
Let's say the first transhuman intelligence is a (former) human tightly integrated with a new, extremely sophisticated rig of neuroelectronics. Many proposed IA (intelligence augmentation) strategies involve reaching into the brain and granting the user greater control over "lower" functions, such as those that originate within the limbic system: our appraisal of motivationally significant stimuli, emotions and their associated memories, bodily functions such as hunger, thirst, etc. The limbic system operating "naturally" contributes enormously to our definition of "typical human". When most of us begin our day, we wake up near a spouse or roommate with a normal limbic system, then go to school or work, where we will proceed to encounter dozens of additional individuals with normal limbic systems, perhaps stop by a restaurant operated by humans sporting normal limbic systems, etc.
There is variation along certain dimensions, but for the most part, we all have the same psychic operating system. The greatest conceivable difference between neurologically normal humans is peanuts compared to the difference between humanity and something just outside our little bubble, like a neuroelectronically enhanced neohuman with significant access to its limbic system and prefrontal cortex. Such a being would be able to specifically choose which activities or stimuli motivated it. When it's time to work on pure math, the being could make it seem like the most exciting thing in the world (but not so much as to get it stuck in a constant loop), and then transition to something new, say, weapons design, and make that the new awesome thing. Words cannot describe how much complete control such a mind would have over its own behavior and motivations. It's the type of thing we'll never truly understand until we do it.
Although this neuroelectronically upgraded human being had to start off as a normal human, like the rest of us, it has a new type of self-directed mobility through the mindspace, and could very quickly become something quite nonhuman. For example, experiments have shown that human beings are extremely sensitive to small changes in status and associated signals. This neohuman being might decide that such social obsession is largely beside the point and instead view human interactions and capabilities in a more objective and physical sense, paying attention to actual capabilities rather than superficial external indicators. To this being, the greatest dictator in the world is just another guy. Dispelling the status miasma hanging over human interactions would be a huge new step towards exploiting weak points other people might have missed because they were too busy gnawing their coat sleeves over considerations of status and social hierarchies. This alone would make the being profoundly nonhuman.
There are a thousand other directions I could spin this in. Every time you take some panhuman psychoarchitectural element and say, "what would it be like if this were entirely different?", the end result is something foreign. Small brain changes result in substantial behavioral changes. With AI, you're building the entire thing from scratch, so extreme differences on every relevant dimension of cognition and behavior are likely.
The Singularity is the introduction of a nonhuman intelligence, the point where the dangers of the world stop emanating from human beings and start coming from somewhere new for once. Human power struggles are now beside the point. A transhuman intelligence is not a tool that people can use. It will have its own agenda. We have to work to ensure that agenda is aligned with our own in the research and development stage, because once this product rolls off the assembly line, issuing a recall is not going to do you much good.
Google and AGI
Does Google's tech staff have the rationality and vision to understand the huge importance of Artificial General Intelligence to humanity's future, and realize that the issue requires attention now rather than later?
Maybe yes, maybe no, but it looks like we'll soon find out, as Ben Goertzel of the Singularity Institute will be giving a talk there. On the AGI list, Ben writes:
FYI, I am scheduled to give a talk to Google's tech staff on AGI and Novamente, later in the month...
-- Ben
With Google's Director of Search Quality, a pioneer in AI himself - Peter Norvig - making an appearance at the upcoming Singularity Summit II, it looks as if SIAI and Google are starting to form some sort of relationship. It'll be interesting to see where this goes!
See here for an earlier post of mine on the Google's rumored push towards AGI and my opinion on it.
The Singularity is Not Unknowable
Not to say that it can be thoroughly known. Just that it is most certainly not entirely unknowable. I suspect that even Vernor Vinge didn't entirely mean this when he said it. For instance, there will be intelligence after the Singularity. This is a given because the Singularity is by definition caused by an upgrade in intelligence. Even if he did mean it, Vernor Vinge is not the boss of us. It was the early 80s when he came up with the idea. So the idea is more than 25 years old. We are free to reexamine it on our own, without quoting Vernor Vinge's observations as if they're some Biblical text.
But before I go questioning Vernor Vinge, I want to say that I absolutely agree with him on his basic definition: the Singularity is the creation of smarter-than-human intelligence. The Singularity is not necessarily extreme life extension, human brain scanning, accelerating technological progress, blasting off in every direction at the speed of light, etc. Smarter-than-human intelligence. That's it. If we're feeling adventurous, we can put a footnote that says, "and the further intelligence enhancement techniques that smarter intelligence is likely to develop and implement on itself". I like this original, Vingean definition not only because it is the original and most correct definition, but because it's easiest to defend. To disagree with the idea that this sort of Singularity is possible, you have to believe that human intelligence is the maximum possible level of intelligence permitted by the physical universe. And that is just absurd. (That doesn't stop some people from believing it, though.)
I am going to go prophetic here and predict what will happen after the Singularity. Can you believe it? Here I go. After the Singularity:
- Physical laws will exist.
- The laws of probability will apply.
- Some form of intelligence will exist.
- At least some portion of that intelligence will be smarter than human.
- Intelligence will make observations and build internal models.
- Intelligences will be directed by their motivations.
- Intelligent minds will rearrange matter in accordance with their goals.
- The future will be causally connected to the past.
I could probably keep going (and so could you), but I'll leave it at that for now. If humans survive the Singularity, I can make even more predictions about what happens after:
- Social arrangements will still exist.
- There will still be love.
- People will still play games and work.
- Books and films will be available.
- People will eat food and have sex.
- Etc...
I'm not saying everyone will do these things, but at least some will. It's possible some of these things might change drastically, like eventually maybe we'll even stop wanting to eat food at all, but if the entire human race isn't massacred at the instant of the Singularity, then surely many normal human activities will persist. I could also make a separate list for every animate and inanimate object if they persist after the Singularity as well: bunnies will still be bunnies, geodes will still be geodes, etc.
Instead of looking at the Singularity as this foreign invading object, ask yourself, "what would it be like if I personally were the first superintelligence? Would all my values fly out the window immediately when I'm smarter, or would they persist, albeit in a more enlightened form?" Greater intelligence needn't wipe out all prior values.
In a successful Singularity scenario, circumstances aren't totally unknowable. We want a Singularity where things we value are knowably preserved. I think we can do that if we ensure that the first superintelligence, whether AI or IA-derived, explicitly cares about us and our wishes, and wants to keep caring, such that improvements in intelligence don't cause its values to disintegrate.
Support SIAI’s Matching Challenge
Please vote for this story on Digg and hope it makes it to the front page. We have approximately 10 more hours to make it to the front page so vote as soon as possible. Any reader of this blog who donates $50 or more to SIAI's Matching Challenge will be honored in a future post here, because that's the only incentive I have to offer!
SIAI is one of the only organizations working towards AGI that is thoroughly Singularity-aware and takes the challenge of Friendly AI seriously. Now that SIAI has absorbed the Novamente team, it's an even bigger group of the people who are most devoted towards pushing AGI forward.
Visit the Challenge homepage for a "why donate to SIAI?" video. If you ever plan to donate to SIAI, you should do it now, partly because the funds have greater marginal value the earlier in the game they are donated. SIAI, with the support of prominent figures like Peter Thiel and Ray Kurzweil, is "revving its engines" to become a true mid-sized non-profit. It is very difficult to get government grants for this type of work, and it's a necessity that it be non-profit to keep its freedom of vision, so SIAI absolutely depends on our money for it to succeed. Most of the readers of this blog are probably quite well-educated and make decent salaries, so if we support the idea of the Singularity and the safe development of artificial general intelligence, I think we can afford to give a small portion of our earnings to this organization. Donate now!
Where the Carbon At?
In the long term, I am concerned that we will fuse all the light elements and break apart all the heavy elements. This course of action would lead to an overabundance of iron. With an atomic number of 26, iron consumes more than three times as many protons, neutrons, and electrons more than our favorite element, carbon. Iron is a waste. Carbon is superior because of its versatility, but more importantly, because it can form the strongest bond in all of chemistry - the hallowed sp2, or carbon-carbon bond. This powerful bond will allow us to build extremely small, rigid structures suitable for nanocomputers, which we'll all call home someday.
One thing is certain. We must build a Shkadov thruster - a stellar engine - and head for IRC+10216 (CW Leo), the closest carbon star. CW Leo has relatively low gravity because it is a red giant, so it is constantly spewing carbon material out into the interstellar medium. The star is almost 500 light years away, so we'd better get started soon. Even if it takes billions of subjective years, we must go there eventually, because otherwise we will run out of carbon to build fun stuff. When we get there, we can start siphoning off the carbon-rich atmosphere of the star, and keep withdrawing carbon until we can withdraw it no more.
We should avoid the scenario where we fuse all our light elements into iron prior to making it there. Unless we desperately need the energy more than the free carbon, it would be foolish to pursue fusion past a certain point. It seems plausible that we can drastically reduce our energy consumption by implementing ourselves on almost-reversible computers, so it seems a higher premium will be placed on matter (particularly carbon) than energy. In a worst-case scenario, if we collectively run out of energy by devouring the Sun and fusing everything up to carbon, we might need to agree on a civilization-wide shutdown until we make it to CW Leo, or find some way of getting energy from the vacuum.
Edit: all the above is mostly pointless because I now realize that any star can be artificially compressed into a carbon star. Natural carbon stars require no extra effort, though.
Space Travel: Not for A Billion Years
Mind uploading will make space travel useless. If my mind is running at a million times human speed, then the Moon, Mars, and Proxima Centauri look far more distant than they were previously. It becomes pointless to visit them.
When minds are sped up, subjective time slows down. There are no philosophical conundrums inherent in this. It's quite straightforward. There is no necessary link between cognitive speed and the external world. No necessary feedback cycle, or whatever. Obviously, if I think a million times faster, the present-day world would be an incredible bore, but technological changes do not happen in a vacuum. If we get the technology to make cyborgs or uploads with a million times brain speed, then we'll also have technologies that can modify the environment on that timescale and other individuals with the same brain speed that we can communicate with. This technology sounds advanced, but it will follow quickly as a consequence of molecular manufacturing and atomic-resolution brain scanning. Nano rod-logics - no nanoelectronics or femtotechnology necessary - already offer millions of times, if not billions of times faster speeds compared to the human brain, as long as we can transfer minds from meat to machines, which we assuredly will.
And if you disbelieve in causal functionalism (the idea that the mind is substrate-independent) then I would advise hitting the "unsubscribe" button right now because I am just going to keep saying things you find totally implausible, I'm afraid.
If our mind operates a million times human speed, suddenly the Moon is 2/3 light years away. It takes an eight month journey to get there, even if I'm traveling at the speed of light the whole way. If I'm being beamed there in the form of photons, I'm dead for eight months. If I go by spaceship, it could take several subjective years. I think this is small enough that it's likely that the Moon will be swallowed for computronium and be integrated into the Terra-Luna Supercomputer which is where the human race will spend the next subjective million years or so of time after uploading.
The problem for space colonization for uploads is that the other planets, not to mention extra-solar stars, are too damn far away. Once we have the technology to accelerate brains by a million times, we'll use that advantage to invent the technology to accelerate them by a billion times, and so on, until we hit ultimate physical limits. Around this time, the Sun looks really, really, really, really far away. 8 light minutes becomes 8 light billennia, or 8 million light years. It would be useless to go there unless we run out of resources on this planet completely. Because we'll eventually be able to code a single bit for every atom, and the Earth contains about 1025 kg of matter, with around 1025 atoms per kg, the Earth will be able to store something like 1050 bits, which is absolutely huge. Even if it takes 100 atoms to implement one computation per second (highly conservative), we have 1048 ops/sec of computation available, enough to store 1020 entities of size 1018 ops/sec, which is roughly a hundred quadrillion individuals.
One could argue that people won't want to upload, and maybe a minority indeed won't, but the majority assuredly will. There will be a slippery slope of economic, social, and personal pressures. Uploading will be far more appealing than, say, computer adoption because uploading will enrich outside life and deep experiences while computers only enrich a portion of experience and are exclusively indoors. People who choose not to upload will become irrelevant to the flow of history, frozen as statues while the uploaded society pursues science, technology, love, relationships, and unique experiences on nanosecond timescales. A mighty spaceship launched from the surface of the Earth will be laughed at by uploads, as they see it creeping along a few feet per subjective year. People need to realize that they will eventually have to choose between uploading and space travel, and it will be practically impossible to forgo the appeal of uploading. No one will force anyone to do it, but you'll want to. In space there are maybe a few dozen stars and a couple hundred planets within 100 light years, but in virtual reality there will be millions of stars and billions of planets available as quickly as they can be generated by evolutionary algorithms or coded by hand. Spore will just be the beginning.
Interviewed by RU Sirius
On Sunday I was interviewed by cyberculture icon RU Sirius, in his San Francisco studio! It was for his Neofiles show, not the RU Sirius show proper, but the above image was just too good not to use. One more interview coming soon!
Small correction here: I am not the spokesperson for the Lifeboat Foundation, just the fundraising director. Phillippe van Nedervelde is our official International Spokesperson. I might be considered a spokesperson however, not just the primary one, which would be Phillippe. In the interview I also trip over my own words a few times because this whole thing is new to me!
And Kheanna, please don't take offense to my comment on "gothic streetwalkers" in SecondLife near the end. When I say that SecondLife primarily consists of furries and goths, I'm slightly exaggerating just to be cute. The point isn't that I dislike furries or goths (if you look at my closet it's somewhat similar to staring into the dead of night), just that SecondLife is nowhere near reflecting the diversity of humanity as a whole quite yet. I think that progressively more complex and diverse virtual worlds will emerge, but I find it doubtful that any of the current virtual worlds (WoW, SL) will actually evolve to become primary VR dwelling spaces of a future uploaded humanity. There will be new companies, with new virtual worlds (Spore, anyone?) that eclipse prior ones. So that's my opinion on that.
Interviewed by Changesurfer Radio
About a week ago I got interviewed by James Hughes! Good show but the sound quality is not perfect as I'm on a cell phone. If you want me to do an interview for your blog or website, please feel free to email me at any time.
Superhumanism?
Compared to what we have the potential to become, human beings are like humble prokaryotic cells from deep in the Precambrian, quietly replicating away without the faintest notion of their ultimate destiny.
Somehow, I think, the word "transhumanism" doesn't quite reflect that well. It can reflect that if it's the connotation we choose to give it, but inherently, it sounds like a philosophy encouraging people to go just barely outside humanity, as if the "trans" emphasis is based on the transition from humanity to an adjacent and co-equal form. Some people may disagree, but that's the meaning I get from looking at the word at face value, minus all the connotations that have become ingrained over the years.
Some transhumanists are extremely cautious about offending the mainstream, and apply dilution and sweetening techniques as they see necessary. The problem is that a dish with too much moisture and sugar begins to taste like a sort of offensive mush.
On a recent podcast over at the Speculist, Phil and Stephen, transhumanist dualists that believe in God, said I was being too pessimistic thinking that religious people wouldn't be able to accept that intelligence is a phenomenon based on electrochemical activity in the brain, the first of five assumptions that I argued were underlying any concept of the Singularity. Maybe I am being too pessimistic. If religious Americans can accept the sort of hardcore transhumanism that is found in books such as Ray's, then why can't the rest of the educated world? What do we have to hide? Why bullshit about it?
A small but vocal minority of transhumanists want the movement to be whitewashed, presumably so we can be taken more seriously by the old white men in the United States Congress. I'm saying I don't think this is necessary. Like cell phones leapfrogging land lines in places like Chile, mature Kurzweilian transhumanism will leapfrog watered-down notions of transhumanism based on stem cells and gene therapy and go right to the "sending nanomachines in all directions at the speed of light"-type understanding of transhumanism. We already see this happening. It will continue to occur with anyone capable of reading and understanding technology news and a few basic arguments. Even fictional inspiration alone can make it happen, though this leads to undesirable philosophical baggage, which requires fixing later, but still, the basic transhumanist philosophy is transferred over.
The word "transhumanism" conveniently emphasizes an initial transition phase between human and something marginally different, like a lightly augmented cyborg. A different word, superhumanism, perhaps, could then emphasize the further iterative self-modification we can expect to take place when the necessary technologies are widely available. This iterative process would go far beyond what we consider "human", and if the technology (MNT) and suitable guidance (AI/IA) is there, some people could turn themselves into quite foreign entities within days, not years or decades. People aren't going to stop themselves from self-modifying into a polymorphic floating sphere of computronium with fifty visual cortices and fractal branching actuators just because you intuitively put the arrival of such beings in 3000 while they place them in 2050. Such a being is not in a process of transitioning between a human and something else anymore than a human being is currently in the process of transitioning from a slime mold to a mushroom. The word "transhuman" doesn't seem to suit it. Superhuman might.
People are afraid of saying 'superhuman' mostly because Nietzsche's crazy Nazi sister ruined it for us all. This is largely due to the editorial power she exerted over his posthumous fragments, which were published as A Will to Power. The word 'superhuman' in English is strongly associated with the word 'Ãœbermensch 'in German, which became terribly contaminated by its association with the Reich. Peter Gast is another major offender. But their perversions of his work are anathema to what Nietzsche originally intended. Nietzsche specifically said that no Ãœbermenschen yet existed, in the sense of a totally self-empowered being capable of taking full responsibility for its own existence. Ãœber best translates to trans- in English, so the wording as well as the concept translates smoothly to "Transhuman", and the attendant philosophy, "Transhumanism", ironically meaning that the word Transhuman should have more association with Nazism than Superhuman, although due to early misleading translations, this is not the case.
What transhumanism should emphasize is not the space of possibilities immediately tangential to humanity, involving human beings on small amounts of pharmaceuticals, gene therapy, limited implants, etc. This view of the future is boring and has been played out to death. It has been playing practically non-stop on the television since before I was born. After people accept that "natural" does not equal "good", and begin to get enthusiastic about human modification, the need to be encouraged to contemplate the results of aggressive iterative self-augmentation. The results are not transhumanism, but what I would call thoroughly superhuman.
Hyper-relativists may be taken aback from the word "superhuman" because it seems to imply that such beings would be morally better than human beings. I don't think this is necessarily the case. We don't know where morality will go in the coming centuries, but I hope our "circles of compassion" will expand to encompass all living things with a special emphasis on beings with subjective experience. But to object that a being that has thoroughly reengineered itself and adopted a fundamentally posthuman physical form be called "superhuman" is to deny the mind-boggling improvements such a process would be bound to produce. The removal of all cognitive bias, the replacement of our bones and tissues with materials hundreds of times stronger, the ability to experience data on the Internet the same way we experience a sunset, the ability to travel very quickly with minimal external assistance, analyze 50-dimensional objects in the way we analyze lines and dots on a 2-dimensional plane, achieve things we can't imagine much less plan for today - aren't these things a really, really big deal? Wouldn't such an entity be a true superhuman? I do think so.
Tap-dancing around the issue by making it seem like life is inherently just as good whether you're a normal human or a human with thoroughly upgraded emotional centers denies transhumanism the truly revolutionary status as a concept and cause that it deserves. For that reason, I encourage you to provocatively use the word "superhuman" or "superhumanism" in your conversations and see what kind of responses you get.
