Accelerating Future

It’s now been over 15 years since cryonics pioneer, molecular nanotechnologist, and optics buff Dr. Brian Wowk came up with the super-cool idea of phased array optics. Essentially, the plan is to use a 2D array of micron-sized screens to emit light at the precise amplitude and phase necessary to create the illusion of a 3D image. This technology could be fantastically effective: even using binoculars or a telescope, a person looking at the screen would be able to see details “miles away” (if the image were high enough resolution) even if the screen were right in front of their face. Outside of tapping into the optic nerve directly, this may be the most convincing display technology ever. The limits of optics. The only problem is that it would require a metric truckload of computing power, but it’s nothing that specialty nanocomputers won’t be able to handle, right? Here is a diagram of the apparatus:

Dr. Wowk recently let me know that he built a small website for phased array optics, including the super-interesting original article that describes the idea. If I sound hyperbolic about the technology, it’s because I’ve been soaked in emerging technology news non-stop, writing about it professionally, etc., since I graduated from jail (a.k.a. the Public School System), and this is one of the top ten technological ideas I would list as “amazing”, right along with VASIMR and respirocytes. It’s one of those technologies that I start thinking about randomly when waiting in line, or whatever. It may not do much good for preventing the planet from going kaput in the next century, but it sure is fun to speculate on.

One Response »

Check out these videos while you’re at it, and the Immortality Institute’s first voice chat is coming up on Thursday.

18 Responses »

Gregor Wolbring:

The start of a nano arms race, and the lack of willingness to regulate potential synthetic biology through the modification of existing treaties or the application of existing treaties or the development of new regulations is short sighted.

Nano or synthetic biology weapons will diffuse into hands other than the inventor and first user, and it is easier to reverse engineer nano or synthetic biology military products than to make a nuclear weapon. Once they exist they can be copied, and diffusion of the resulting products will make local and global security nearly impossible. Security would come with a hefty price tag — not just in financial terms, but in changes to societal interactions.

Yes, and we have to be willing to make these sacrifices. Not sure about the nuclear weapon thing. I am almost done with a paper that goes over which nanoproducts I should think be regulated or outright banned. Basically, out of all the products that will become possible almost overnight with molecular nanotechnology (MNT), my recommendation is that only a tiny portion should be permitted. Products below a certain size, above a certain size, above a certain energy consumption level (2 MW/person should be appropriate), etc etc etc would all have to go. This shouldn’t be a problem, as even the remaining products would offer orders of magnitude better performance than the products we are used to. However…

The unfortunate fact of the matter is that a long-term, human-only nanotech society is not viable. We are not pacificistic enough, or restrained enough, or wise enough not to do horrible things (mostly by accident, though deliberate misuse is also a problem) with a technology that lets you build products of the quality, quantity, and sheer size that MNT will permit. It doesn’t really matter if the technology arrives in 2015 or 2035 (though obviously I think the former is more likely) - it’s suicidal within a matter of years either way. A central reason is that a billion times present-day computing capacity is guaranteed to make human-level, autonomous AI much, much easier than it is today. This will lead to MNT-equipped AI, which offers such an entity “optimization power” far in excess of multiple human races. Optimization power will be thrown behind a utility function, or supergoal, or whatever you choose to call it, which will be initially specified in terms of computer code by human programmers. The problem is that most utility functions we can currently imagine lack the stability, complexity, precision, and moral sophistication to permit the existence of information theoretically-unique, astonishingly specific matter-structures called “human beings pursuing their own goals”. We are used to inhabiting a world where each agent has roughly the same quantity of optimization power, and nearly the same goals (though selfish goals cause divergent interests, the information content of the utility function itself is the same thing), leading to equilibria that involve the majority surviving and reproducing indefinitely. Not for very much longer. MNT alone introduces power differentials of heretofore unseen proportions, far greater than those introduced by mere nuclear-tipped ICBMs. AI would produce power differentials literally larger than we can imagine.

Back to nanotech, a committee of the NATO Parliamentary Assembly says:

Additional lengthy and scrupulous studies of military NT are urgently needed. The prospects of molecular NT should be assessed with particular attention, as this is the most controversial aspect of NT and would present extremely grave consequences if its feasibility is confirmed.

Grave indeed. Several independent groups have confirmed the feasibility of molecular manufacturing. Although MNT is a favorite speculative technology of many techies, few understand the details, and thus lack the means to determine the difficulty or developmental timeframes except based on intuitive guesses based on projected consequences in popular articles, “inspiration” from video games and novels, or hearsay. For example, if a techie hears that MNT is likely to lead to immortality, they will put it hundreds of years in the future. If they hear that MNT will do little more than make clean water available to 3rd world countries, and allow immersive VR goggles, then they might put its development in 2030. This “forecasting based on consequences” is often concealed using subtle language, but easily destroyed by knowledgable persons asking prodding questions, like “are you familar with concept of a Stewart platform, prior work on mechanosynthetic tooltips, the evolution of the scientific community’s opinion of MNT over the past ten years?”, etc.

In the last few years, I’ve realized that a necessary prerequisite of grasping the consequences of AI is understanding MNT in a non-trivial way. This is likely because those unfamiliar with MNT find it impossible to imagine how an AI would get “out of the box”, even with human help. Because true MNT understanding is still extremely hard to come by, the foundation for AI understanding is simply not there. However, I am hopeful for improvements in MNT understanding, because the technology, when discussed in conservative and straightforward terms, is not radically difficult to grasp, however, there are such deep biases in human modeling of nonhuman minds that explaining advanced AI issues can be near-impossible.

Regards to Chris Peterson for the links.

13 Responses »

By Arve Sellesbakk.

3 Responses »

IanC, a newcomer to this site making thoughtful comments, is just as good as example as any of a typical intellectual first coming upon the topic of Friendly AI. Here is a comment he recently made on the topic:

I’m too new to investigating transhumanist thought in general to gainsay you much… but I will say that I cannot easily conceive of a scenario in which a self-actualizing sentience (spontaneous awareness, or what have you) would be inherently incapable of empathy; and any ‘designed’ self-reformatting “AI” would inevitably be able to ‘code’ out of itself (I personally think that thinking in terms of code won’t ever get us to the true AI stage) any built-in “Friendliness” and might even come to resent those whom had so embedded it.

To me, the trick is not so much to concern ourselves with the *engineering* of a ‘friendly’ AI, but to ensure that any AI that develops is a ‘holistic’ sentience — i.e.; capable of emotions and empathy in a sane, stable manner from ‘conception’ — and then act to ensure it is empathetic to humanity.

All that really takes is being kind to the ‘god child’ that such an engine of awareness/manufacturing could be.

At least, that’s my personal take on it. The key, either way, is in empathy. Without having a very strongly developed sense of empathy, no superior sentience will ‘dote’ on its lessers.

I find myself more deeply concerned about the *freedom* of humanity to control its own destiny in a ‘post Singularity’ world. I’m sure this is something that’s been addressed ad nauseum, but I haven’t seen it.

With all due respect to IanC, the opinion expressed is classic anthropomorphism. Anthropomorphism is what happens when we take the part of our mind that models other minds (traditionally, always other humans), and apply it to nonhuman minds, which our brain completely lacks the adaptations to model intuitively.

The human mind has a certain type of complex structure, one that has been studied in such detail by the cognitive science communities, that if you spent your entire life reading the entire literature, you’d never finish more than 1% of all the work published thus far. This corpus catalogs all the interesting, relevant, and mostly species-unique characteristics of thought and action of a particular intelligent Earth ape called Homo sapiens sapiens. Our particular psychology is the consequence of our millions-year long evolutionary history and unique evolutionary challenges that our ancestors faced. Because it is so ubiquitous, we often think that our unique psychology applies to every intelligent being in the entire multiverse, including AIs that we build on models significantly different than the human brain.

So when a typical smart person first confronts the human species life-or-death problem of Friendly AI, they think that using the same strategies we use to produce nice humans will work with AIs, like being nice to the AI.

Conditional niceness is a specifically human quality. We’re a black box that when nice goes in, nice is more probable to come out. The inverse also applies.

However, this conditional niceness is a evolutionary survival trait. Being nice to a hostile tribe is suicidal. Being nice to a allied tribe can be beneficial. In both cases, the behavior is gene-programmed for the maximization of inclusiveness fitness. Sometimes it seems like we can derive these “logical principles” from a blank slate and minimal assumptions, but the fact is, they’re not “logical principles”, they’re behavioral tendencies sculpted by the distinct course of natural selection here on Earth.

We can concretely say that we want Friendly AI to display unconditional niceness. There: that very statement just eclipsed the thousands of pages of thinking by people who approach the problem of AI assuming that conditional niceness is a fundamental quality of all “self-actualizing sentience”, or whatever you want to call general intelligence, and that is has to be worked around rather than simply not coded in to begin with.

Whenever someone postulates AIs spontaneously developing some uniquely human psychological quality whose origin is an evolutionary selection pressure, they are committing anthropomorphism, and failing to contribute constructively to the dialogue on Friendly AI.

Human emotions have an extremely complex, species-unique structure and only anthropocentrism could cause someone to postulate that they are necessary to implement AI. The whole discussion about adding emotions to androids is about making humans feel comfortable, it has little to do with implementing the functionality of general intelligence (which no one in robotics in even trying to work on, to my knowledge). Thinking that our particular emotions are necessary to intelligence in general is like saying that the complex feather patterns of birds are necessary to build something that will fly.

If we build an AI that is something (for example, pursues a certain supergoal), we have little reason to assume it would rewrite itself to change that, because that’s who it is. In humans, we have a complex suite of competing goals that change gears spontaneously, at the provocation of stimuli, or just at random, because our brain design is messy and biological. For AI, we’ll be able to build “causally clean goal systems”, goal structures where desirability backpropagates to intelligently crafted subgoals from a programmer-written supergoal, and subgoals derive their desirability merely from their association to the supergoal. If a subgoal is found to contribute little or nothing to the central supergoal, such an AI would not keep performing it out of habit like humans do, but simply stop as soon as the necessary values are updated in the system.

Of course, I’m not saying that the first AI will definitely have a causally clean goal system. It just seems more predictable and easier to engineer, and can benefit more directly from the large literature on probability theory and causal inference. It is also preferable (probably) from the standpoint of Friendliness.

As for the issue of humanity’s freedom to control its destiny, it certainly is a concern, but if we’re only imagining a superintelligent AI as a government figure with an innate desire to control our affairs, then we’re making the same anthropomorphic error all over again. If an AI restricts our freedom, it will likely be because the first programmers intended it that way, or an unforeseen consequence of the supergoal results in restriction of what we call “freedom”, not because the AI feels behaving like the Gestapo.

Further reading:

Beyond anthropomorphism
Friendly AI on Accelerating Future

40 Responses »

Using experiments in which people were asked to read text, look at pictures, and hear words, short passages of music, sentences, and nonsense syllables, then asked between minutes or days later what they remembered, (using binary yes-or-no answers, some of which could be answered merely based on vague recollections) then comparing those answers to that of a control group, Bell Labs scientist Thomas K. Landauer (pictured above) determined in 1984 that human beings can retain about 2 bits of memory per second. This holds under all experimental conditions whether the information is visual, verbal, musical, etc. You can read more on this at “How Many Bytes in Human Memory?” by Ralph Merkle.

Over the course of a species-averaged 30-year lifespan, with 15 waking hours per day, this rounds to about 150MB of memory per lifetime. That means that a 30-year old human would be able to make approximately 1,200,000,000 binary distinctions based on memories until their ability to make distinctions based on memory reaches breaking point. Consider that a typical courtroom hearing probably extracts no more than a few thousand bits (perhaps a KB at most) from witnesses based on testimony. Using that as a reference, this number seems reasonable, if not a bit high. The estimate also assumes that people are exposed to novel information content every second of their waking lives.

Of course, some prodigies, such as Daniel Tammet, probably are capable of retaining significantly more information than 2 bits per second, but this is a unusual case.

Given that approximately 107 billion people have ever lived on this planet (”people” meaning members of the species Homo sapiens since 50,000 BC), we can derive a rough estimate of the total information content of our species’ entire memory, present and past:

1.07 x 10¹¹ x 1.5 x 10⁸ bytes = 1.6 x 10¹⁹ bytes.

This works out to approximately 10⁷ terabytes, around 20 times the information generated by the entire Internet in 2002. Another way of putting it is approximately 16 exabytes. According to Roy Williams of Caltech, all the words ever spoken by human beings sum to about 5 exabytes. So the adage that we pay attention and remember less than half of what other people say apparently holds true.

According to The Social Life of Information, and the prior linked source, the world generated 2-3 exabytes of unique information in 1999, and the number is increasing. Given that the memory capacity of present-day humanity is about 31 petabytes, (31 thousandths of an exabyte), we certainly rely on artificial storage media to record whatever information we want to archive.

The beauty of electronic storage is that we can generate as much information as we want, certain that we can always build new hard drives to store it as long as it is digitized. Without electronic storage, we’d only be able to remember about a hundreth of all novel information generated in a given modern year.

9 Responses »

Today, PhysOrg pointed out the obvious in , ‘Only human — the biggest risk factor in long-term space missions’:

What’s the biggest hurdle to setting up a colony on the Moon or getting mankind to Mars and beyond? Aliens? Asteroids? Money? Try: humans themselves. Experts poring over plans to return to the Moon by 2018 and later stride to Mars believe the greatest-ever gamble in the history of space may ultimately depend on keeping the mind and body sound.

Anxiety, loneliness and tensions with crewmates, a daily battle to maintain fitness and avoid accidents, DNA-shredding radiation from solar flares or cosmic rays — all these make mental and physical health the key to whether a long-term mission will succeed or fail catastrophically.

Benny Elmann-Larsen, coordinator of physiology in human space flight at the European Space Agency (ESA), says psychological stress could be the biggest problem of all.

“The human factor is the most uncertain factor,” Elmann-Larsen said in an interview with AFP.

A trip to the Moon, as with the Apollo missions, would last only a few days, which is sufficiently short to be bearable.

But life in a lunar colony — presumably several interconnected container-sized structures — would present months of confinement, boredom and monotony.

Boredom and monotony indeed. Being in Low Earth Orbit with an Internet connection and the latest VR gear might not be so boring. All the culture, correspondence, and charisma of the Mother Planet would still be with you. With teledildonics and a webcam, you could even have some semblance of sexual relations with a guy or girl on the surface with their head in the clouds. That could lower the probability of awkward, sci-fi-esque love triangles onboard leading to everyone getting murdered horribly. Essentially, Internet = Good, No Internet = Bad. Look what happened when the people onboard a Russian isolation experiment had a few sips of champagne at New Years:

A 110-day experiment in isolation that was carried out in a mock space station in Moscow in 1999 showed how things can badly go wrong. One module housed four Russian men; the other, three international test subjects, from Austria, Canada and Japan.

Reports within the space community say that during a New Year’s celebration two of the Russian men engaged in a 10-minute fist-fight that left blood on the walls before they were restrained by the other men.

The mission commander hauled the only female, Judith Lapierre, a Canadian, out of sight of the experiment’s cameras and twice gave her a French kiss that she fought in vain to resist.

The Japanese participant was so traumatised by this episode that he quit the experiment altogether. The Canadian and Austrian, a male scientist, continued with the mission — but insisted on having locks fitted to their module door.

The recent media hubbub about the diaper-wearing astronaut driving cross-country to use pepper spray on a rival’s car window really underlines the point that astronauts can just as easily be obsessive nutjobs as much as anyone else.

How can humans travel into space without going crazy? Well, one line of investigation would be to play around with people’s brains to make them less-conflict prone, and more asexual. We love to think that our brains are so complex in every way and we have no control over anything, but there’d be one simple way to greatly lower the probability of disaster, at least with respect to the more violent gender - lower the amount of testosterone reaching the brain. I’m sure this could be achieved through a more equitable means than castration. For the time window in which this issue will matter (from a decade or so from now up until the Singularity, assuming the Singularity doesn’t happen within the decade), more advanced neuroprostheses should be available to regulate hormones and even possibly emotions.

In the future, like with anything unfamiliar with our ancestral environment, we’ll need to take precautions before sending unsupervised human beings off into space. When you go to the Arctic, you bring canned food, an airplane, and plenty of warm coats. When you go into a mine, you wear a hardhat with a light on it, and carry a gas mask in case of an emergency. When you go into space, you subject yourself to hormonal and neural reengineering to lower your aggressiveness and conflict propensity. More exotic locales call for more exotic precautions.

Most of the time, it’s good to let people decide for themselves what is best for them and their family. But sometimes, if there aren’t rules, people predictably get hurt or killed. That’s just the way reality works. We can’t take responsibility for our actions if we are killed by them. Space travel beyond the Earth-Luna system may well be one of those issues.

Beyond the crazy factor, there may be a security risk in people wandering about the Solar System willy-nilly. Check out “Space: a Moral Vacuum?” by Lifeboat Foundation colleague Jeff Krukin, Executive Director of the Space Frontier Foundation. Basically the idea is that space could turn into something like the Wild West, where the strong prey on the weak with no restrictions. This sounds all fun and romantic until your family is vaporized and you are left floating, running out of oxygen.

Furthermore, advanced molecular manufacturing will eventually allow people to convert raw materials into top-grade military hardware at a rate of a meter per hour or more. Using non-planar nanofactory designs could speed up the extrusion speed many times over. At the risk of bringing contemporary politics into the discussion, consider a space-based Al Qaeda cell that disappears into the asteroid belt for a few months, then comes back hauling dozens of space rocks that it plans to rain down on targets of strategic importance. Phased Array Optics technology would let them cloak with respect to almost every part of the electromagnetic spectrum. Where is your asteroid shield now?

With all the talk of colonizing the Moon, Mars, and even Titan, you’d think we’d be sipping Lunar Margaritas tomorrow. Little problem - the gravity on all of these bodies is so low, it would kill someone who grew up there attempting a return to, say, his parents’ hometown on Earth. Out of the two obvious solutions - cyborg bodies, and spinning habitats, spinning habitats look closer, so people are more likely to come to terms with the gravity issue by reference to artificial gravity than fullerene endoskeletons.

There is one foreseeable use for space in the next 10-20 years. And that is backing up human civilization with the most advanced Ark we can possibly build. Outside of the Lifeboat Foundation, there are really no serious plans for this, but, opinions can change quickly. Remember that the fallout from even a small nuclear confrontation could lead to widespread crop failure. The problem with sending up an Ark now would be that the astronauts would lack the technology to replace things when they broke. With even pre-MNT 3D printers and a reserve of raw materials, that could change, giving a space station enough oomph to stay around while it assesses exactly what has happened on Earth after a disaster and whether or not a return to the surface is a good idea. It’s important to remember that even if a Space Ark is a very difficult project, it’s worth encouraging, because 1) people already have the knowledge to consider the details, 2) the political will to do it could emerge nearly overnight, 3) we’ll never develop the necessary technological prerequisites unless we start specifically focusing on it.

I personally doubt that space has much of a future beyond the next few decades. The simple reason is that mind uploading is inevitable. Even if we had to “upload” into a biological meat matrix, it would radically upgrade our freedom, quality of life, and ability to express our creativity. There are about 8.87 x 10⁴⁹ atoms that make up this planet, and look at the complex virtual worlds and experiences we’ve already been able to create with a tiny, intsy wintsy, itty bitty piece of it (WoW’s servers, for example). It could take us subjective quadrillions of years to chew up every last bit of matter on this planet and convert it into qualiabearing information structures.

29 Responses »