The New Tree of Life Monday, Apr 30 2007 

Various futurist thinkers, and even a few unusually insightful journalists have pointed out that the next generation of threats to humanity are self-replicating technologies.

It’s really too bad that self-replicating technologies are so dangerous, because such technologies would also be most useful for completely realigning the material structure of the world with our collective will.

Self-replication is something that life figured out a long time ago. The DNA in every cell of your body has a replication lineage that can be traced, in an unbroken line, all the way back to the first self-replicating piece of genetic material ever. Every branch on the Tree of Life is a descendant of the first self-replicating organic being.

The problem with this mode of organic self-replication we’re so familiar with is that it’s relatively restricted in what it can build. Water is essential in large quantities for all the processes of life so any terran biological life form must be made up of primarily H2O. This requirement strongly constrains all material properties, including freezing point, boiling point, tensile strength, density, toughness, and many others. The Tree of Life is located in a narrow chemical ghetto.

Humanity’s task is to reinvent the technology of self-replication that the first single-celled organisms stumbled upon billions of years ago, but this time, do it right. Expand the range of possible forms. Boost the throughput and decrease the minimum duration of each self-replication cycle. Combine the adaptive elegance of the biosphere with the superior absolute performance and chemical flexibility of the technosphere. The final outcome is something far greater than both: the means to turn the inorganic to the organic and vice versa at arbitrary rates and generate new forms of more diversity and ability than either sphere could generate alone.

The basic concept of a non-biological or super-biological self-replicator dates back to a posthumous work by math/comp-sci superhero John von Neumann. More recently, our friends Robert Freitas and Ralph Merkle did a comprehensive review of everything known about kinematic (physical) self-replicators, including a 137-dimension classification system that subsumes all known and proposed self-replicating machine systems, and presents numerous plans for artificial self-replicators to be built in the near future. Most futurists and “technology experts” are cheerfully clueless about the coming revolution in artificial self-replicators, but not everyone. For example, Gregory Cochran, an adjunct professor of anthropology at the University of Utah known for his work in adaptive optics and evolution of the Ashkenazim, recently came out with a “don’t you people see what’s coming?” statement on Edge.org:

“In the sweat of thy face shalt thou eat bread”—it has always been that way.

Most men have been slaves of necessity, while the few who were not lived by exploiting others who were. Although mechanization has eased that burden in the advanced countries, it is still the case for the majority of the human race. Limited resources (mainly fossil fuels), as well as negative consequences of industrialization such as global warming, have made some people question whether American living standards can ever be extended to most of the human race. They’re pessimists, and they’re wrong.

Hardly anyone seems to realize it, but we’re on the threshold of an era of unbelievable abundance. Within a generation—sooner if we want it enough—we will be able to make a self-replicating machine, first seriously suggested by John von Neumann.

Such a machine would absorb energy through solar cells, eat rock and use the energy and minerals to make copies of itself. Numbers would grow geometrically, and if we manage to design one with a reasonably short replication time—say six months—we could have trillions working for humanity in another generation. You might compare this process to a single cell of blue-green algae, which replicates over the summer until it covers the entire pond. But unlike algae, a self-replicating machine would be programmed and controlled by us. If it could make it its own mechanical and electronic parts, it would also be able to make toasters, refrigerators, and Lamborghinis, as well as the electricity to power them. We could make the deserts bloom, put two cars in every pot, and end world poverty, while simultaneously fighting global warming. It’s closer than you think, since the key technologies are already being developed for use in rapid prototyping and desktop manufacturing. Aristotle thought that slavery would only end when looms weave by themselves: we’re almost there.

This is something that people should get with no problem. We’re on the edge of an era where we can make anything from raw materials for free, and quickly. This is so unlike anything humanity, or indeed life on Earth, has ever experienced, that it warrants sitting down for a couple hours every week and pondering in detail.

With self-replicating workers that get their energy from the Sun and their building material from the Earth’s crust, there is no limit to the scale of engineering projects. We’ll be able to dig holes of such depth and width that there will be a risk of the atmosphere being heated up by portions of exposed mantle. Pollution will be a thing of the past, as long as new self-replicators can recycle the raw material of old self-replicators into new forms, which seems chemically plausible.

Geoengineering - significantly reengineering the structure of the Earth as a whole, will become possible. Advanced technology, including displays, actuators, and devices for communication and computation will be available in unlimited quantities, as long as they require no rare elements. Our current knowledge tells us that such devices can be built using elements fairly common on the Earth, such as iron, carbon, silicon, etc. This will lead to products such as mile-wide displays and completely customized surroundings, similar to those in a virtual environment. Peeling matter from the planet in an onion-like fashion and arranging it in concentric shells, each illuminated by sunlight rerouted from the top level, could provide living space for trillions of human beings, expanding the Earth’s available surface area from 500 million km² to dozens or hundreds of billions of km². The huge quantities of iron available in the planet’s core could be used to create a framework of pillars and levels on which to layer the lighter elements. Layers could be made permeable by installing motion-sensing doors across large areas of their surfaces. Screens covering the bottom of each layer could provide the illusion for those living in the layer below of empty blue skies, complete with a day/night cycle and simulated weather. The layers could be covered in forests or grasslands, providing miles of open space for anyone who wants it.

Yes, the above is speculative, but entirely permitted by the laws of physics, and if the rearrangement of matter on this scale still sounds incredible to you, you need to learn to appreciate the great potential of artificial self-replication. The “Unfolded Earth” scenario I’ve described above may sound too fantastic for some, but it’s just a tiny sample of what will become possible when we unleash the power of self-replication and make it our own. Yes, there are numerous risks, but if we can handle them, the possible rewards go beyond our wildest dreams.

Crossing the Line Friday, Apr 27 2007 

Crossing the line to superintelligence is of vastly greater significance than the invention of fire, the Internet, or landing on the Moon. All these accomplishments came from human-level intelligence. Boost the underlying intelligence itself, and you’ve done something far deeper than create some new external product of human-level intelligence.

For example, consider the world from the viewpoint of a Homo erectus. They had tools - handaxes. These tools were of various types - pointed, cordate, ovate, ficron and bout-coupé shapes, cleavers, retouched flakes, scrapers, and segmental chopping tools. Flint, basalt, chalcedony, quartzite, andesite, sandstone, chert and shale were all used as raw materials to build these axes. Some were very large and probably just ornamental. Some were discus-shaped and possibly used as hunting weapons. It is thought they also had a social role, with enterprising Homo erectuses fashioning better tools for greater peer approval. From the viewpoint of one of these guys, they had command over a remarkable number of handaxe forms and designs, and put them to use for a variety of different purposes.

From the viewpoint of an intelligence smarter than us in the way that we’re smarter than Homo erectus, all our technology, from planes to trains to lamps to sinks to nanotubes to satellites to linear accelerators, probably look just like variants on the same basic handaxe. Our descendants or future selves will not look back on us admiringly, and say, “golly gee, these guys were so clever that no leap in intelligence ever happened that bested the difference between them and their immediate predecessors!” They will not be genuinely impressed with what we are doing, any more than we are genuinely impressed by a pre-Neolithic hand axe. If we were to show them our greatest technological achievements, they might pretend to be genuinely impressed, so as not to hurt our feelings, but really, they’d probably be daydreaming on the side about mechanisms of such complexity that no aggregation of human beings, no matter how numerous or intelligent, could ever make sense of it all.

I believe that a lot of Singularity skepticism derives from people who don’t get that we’re not the highest form of intelligence that the universe permits to exist. Being a computer science poindexter sometimes hurts more than it helps, because such people are accustomed to being the smartest ones in the room, making it all the more difficult to imagine an intelligence that not only blows them out of the water quantitatively, but can think thoughts they can’t think, even in principle. When people say, “oh, we’ll be able to fight the superintelligent AIs with our rebel guerilla group!”, or “we’ll nuke it to smithereens if it disobeys!”, they don’t get that, once it’s smarter than you, you’ve already lost. Once you’re dealing with something genuinely smarter than human, you have to rely on the hope that it doesn’t want to hurt you, not the assumption that your crappy “foolproof safeguards” will do a lick of good against a true superintelligence. Eliezer came up with the AI Box game to help hammer this point into the collective consciousness.

This is why I raise an eyebrow when people tell me they believe in a slow or incremental AI or IA takeoff. Once you cross the line, you’re quite simply not in Kansas anymore. The scale-up to the point of human-equivalent intelligence may be a slow process, but once you go beyond it, we as humans lose our privilege to say what this new mind can do. Our license to put down limits is permanently revoked. This doesn’t mean that we can’t predict anything - Vinge was wrong when he said that the Singularity is a point of absolute unknowability. If a selfish entity is the first to cross the line, it doesn’t bode as well for humanity than if a selfless and benevolent entity crosses the line. The method of line-crossing (BCI, AI, IA) will influence the play-out of post-line activity, at least until a degree of progress is reached where the post-human point of origin becomes moot. But methods and motivations aside, when we’re talking solely about ability, the prudent assumption to make is that superintelligence is sufficient to shatter most barriers we can conceive of. Self-improving superintelligence all the more so.

Short Article on Human Universals Thursday, Apr 26 2007 

“Human universals” is a term used in anthropology and evolutionary psychology to refer to behavioral or cognitive traits common to all neurologically normal humans. The notion of human universals was partially formulated as a challenge to cultural relativism, a predominant view of human nature in the late 20th century, which some psychologists and anthropologists see as greatly exaggerating the variance among members of the human species.

In a book of the same name published in 1991, professor of anthropology Donald Brown listed hundreds of human universals in an effort to emphasize the fundamental cognitive commonality between members of the human species. Some of these human universals include incest avoidance, territoriality, fear of death, rituals, childcare, pretend play, mourning, food sharing, kin groups, social structure, collective decision making, etiquette, envy, weapons, aesthetics, and many more. Wider recognition of human universals has led to a sort of mini-revolution in psychology, which has begun to take more input from the harder sciences of anthropology and biology, and less from the ubiquitous pop-psychology of the 20th century.

One of the greatest popularizers of the notion of human universals in recent years has been from Steven Pinker, a cognitive scientist at Harvard and author of four widely read books on the human mind. As a champion of the rising science of evolutionary psychology, Pinker argues that, in the same way we all have ten fingers, ten toes, two eyes, two ears, and a mouth, all with the same basic biological features from person to person, we should expect our cognitive features to have similar commonality. The psychological differences between human beings are then differences of degree, not in kind.

The existence of an experimentally verifiable set of human universals has two key consequences. The first is that it makes further psychological experimentation and research more valuable than some may have thought. If we can identify the common cognitive features between us and their characteristics, we learn not only about every human culture and individual on earth today, but of those into the indefinite future, as long as their genomes stay essentially human. The second is that the human species has more in common than conventional psychology would have us think - that conflicts arise in spite of our fundamental cognitive similarities, rather than from them.

Why is AI Dangerous? Thursday, Apr 26 2007 

To put it in a single sentence, I’d say that it’s because only a minority of cognitively possible goal sets place a high priority on the continued survival of human beings and the structures we value.

Another reason is that we can’t specify what we value in enough mathematical detail to transfer it to a new species without a lot of requisite hassle.

It would be easy if we could just transfer over the goal set of a “typical human” or a “nice person” and hope for the best. But there’s a problem: we have no experimental evidence of what happens when a human being can modify its own goals, or increase its own intelligence and/or physical power exponentially.

What little evidence we have of scenarios where people acquire a lot of power in a short amount of time indicates that the outcomes are usually not pretty. In fact, we have complicated democratic mechanisms built into our society to guard against these types of outcomes.

Most AI designers are missing the challenge because no one wants to have to take the responsibility of creating the first truly intelligent being. They just want to play with their program. The idea of taking any responsibility for the products of one’s research is a relatively recent notion, one that only holds weight with a minority of scientists and engineers, even today. This is usually because scientists and engineers are embedded in a large institutional apparatus that places responsibility so far up the chain of command that the actual researchers are absolved of most, if not all responsibility.

Back to the original issue of goal sets. Here are some likely applications for the most advanced AI technologies in the next 10-20 years:

  • Intelligence analysis and wargaming. (link)
  • Law enforcement (link)
  • Analyzing interstate politics (link)
  • Finance, banking, & investing (link)
  • Controlling combat robots (link)
  • Automating work flows (link)

There are many others, but I put these on the top of the list because they have the most economic or political importance, and therefore will be getting the most research money.

As AI in these areas progresses, the systems will go from outputting decisions only when explicitly requested, to outputting decisions continually and automatically. When a human worker consults the machine for input, it will be more like dipping a cup into a stream and tapping into the preexisting flow of knowledge consolidation and decision-making, rather than flicking on a light switch or pressing “run” for a conventional computer program.

Being continuously thinking, continuously decision-making entities, these AI systems will have implicit top goals, whether people explicitly program them or not. The implicit top goal of a workflow automator will be to accelerate the completion of productive tasks. The implicit top goal of the finance bot will be to pick stocks that maximize return on investment. The implicit top goal of the combat robot AIs will be to take out or capture people specified by certain data files in its memory.

What makes AI potentially so dangerous is the lack of background common sense and humanness that we take for granted. When the clock hits 5, most workers put down their tasks and are done for the day. They go home and spend time with their family, watch TV or play games, or just relax. An artificial worker would have no such “background normality” unless we program it in. It’s on task, 24 hours a day, 7 days a week, as long as its computer continues to suck power from the wall.

It’s that kind of monomaniacal devotion that puts humanity at risk from AI when it begins to step out of the lab and into the real world. An AI with implicit top goals will want to reinforce those goals and achieve them more effectively, where the “goals” are not the same as what you’d see in a human that was handed a piece of paper with those goals written on it, but as they are represented in the context of the AI’s decision structure and worldview.

Reasonableness and sensibility about goals are not easy to transfer over to a mind without the knowledge and common sense built into every neurologically normal human being. A blank slate intelligence sitting in the middle of a forest would be able to build models and make inferences about numerous aspects of its surroundings - that trees are tall, that animals are mobile but plants aren’t, that the weather changes in cycles. But inferences about “the right thing to do”? You can’t derive an ought from an is. Putting an AI in a social environment with humans or other AIs doesn’t help, because without some deep-seated motivation to care about this weird “morality” thing in the first place, an AI will just happily go about accomplishing the subtlety-devoid goals it was originally assigned. As it gains the ability to improve on its own intelligence or tap into the power of robotics, it will continue to get better and better at achieving those goals and harder and harder for humans to reach in and grant it the motivation to care about morality in the abstract.

If AIs in any of the applications I listed before gained the ability to improve upon themselves significantly, either mentally or physically, the implicit top goals they were given will be magnified many times over. There would be little reason for the AI to modify those goals unless such flexibility mechanisms were explicitly programmed in. When a human sees someone starving, they tend to feel sorry for them and at the very least wish they could help. When a human sees someone attacking a defenseless child, they tend to get angry. To your typical AI, a person starving or a child being attacked is only relevant in the context of the goals it already has - “how does this starving human affect stock prices?”, or “can this starving human give me information regarding the location of my next target?” are two inquiries that might come to mind.

Freedom, empathy, self-determination, consensus-building, conflict resolution, aesthetics, camaraderie and rapport - these values and inclinations are built in automatically for every human without serious brain defects. For an AI to share them, they have to be put in terms of lines of code and mathematical rigor. What programmer has the time to do all that work when general intelligence without the human-like morality will be significantly easier to achieve?

It’s that difficulty disparity between stripped-down general intelligence and morally-sophisticated general intelligence that makes AI so dangerous in the long term.

The Planet Everyone’s Been Talking About Wednesday, Apr 25 2007 

I don’t quite understand why some people are getting all worked up about the news of a possibly human-hospitable planet 20.4 light years away in Gliese 581.

First, we have a human hospitable planet right here that we’ve barely even begun to use. In a post last September, I outlined how the Earth could easily hold 100 billion people, if not more, by colonizing the deserts and highlands. I didn’t even talk about the oceans, polar regions, underground, or low Earth orbit. To those desperate to get off the planet post haste, I ask: where’s your creativity? Do you realize that we could hollow out regions the size of cities underground, reroute sunlight from the surface down into them, and have a perfectly nice living environment, with none of the inconveniences of outer space such as: lack of organic chemicals, ice cold temperature, ionizing cosmic rays, lack of gravity, lack of water, deadly vacuum, etc? If living underground isn’t your cup of tea, then there is the possibility of ocean colonies, powered by the temperature differential from the deep ocean and the surface, or polar domed colonies, or airship-supplied mountain colonies, or… really, the sky’s the limit, and by sky, I mean space habitats at an altitude of 200km.

Second, even if we did need to leave the Earth, there is a tremendous amount of raw materials for space colonies right next door in the form of carbonaceous asteroids, which make up about 75% of known asteroids. The asteroid belt contains about a million asteroids of 1km diameter, and a great big planetoid (Ceres) about 1000km in diameter, all there for the taking. Some of these asteroids wouldn’t even need to be reprocessed entirely, but could be turned into viable colonies simply by hollowing them out, pumping them full of oxygen, and getting them spinning. The rock is a natural cosmic ray shield. Through exploitation of the resources of our Solar System in this way, we could create colonies for billions of billions of people, if not more.

Third, I submit that we should think carefully before sending off colonists to far-away places without ensuring that they’re capable of protecting the fundamental freedoms of their citizens, and not degenerating into the primitive tribes that humans seem automatically programmed to create in the absence of a checks-and-balances infrastructure. How are we going to make sure that they don’t accidentally create a Blight that the home system is then helpless to deal with? With the matter-energy resources of an entire star system, it would be a Class 1 hassle for us to come up with defenses against a malevolent entity of that category. Even the possibility of such difficulties may make it undesirable for us to expand too fast too soon.

Human Intelligence and the Multiverse Tuesday, Apr 24 2007 

Humans find it hard to imagine intelligences smarter than we are because we’re designed by evolution to ignore the problems we can’t solve and focus on those we can. Doing it any other way would be an inappropriate use of cognitive resources.

What are the top five elements in your body and their relative proportions? You can’t answer? What’s taking you so long? You don’t even know what you’re made of?

Fact is, humans are pretty damn stupid. Not stupid relative to me or stupid relative to Einstein, but stupid in the scheme of things. Stupid relative to what we could be. We can offer any number of excuses, but in the end they’re nothing but excuses.

Homo sapiens evolved out of the primordial muck. We’re what happens when the muck gets just barely smart enough to reflect upon itself and manipulate its environment significantly.

There are two anthropic pressures at play here. Let’s assume, like Max Tegmark and other physicists, that we live in a gigantic multiverse where all possibilities are realized. The sector of the multiverse capable of harboring intelligent life, or life of any type, is extremely small. If our spatial dimensionality were different, or the intensity of the strong force, or the fine structure constant, or any number of other fundamental constants varied by even a tiny bit, life in this universe would be impossible. Tipler and Barrow beat this point into the ground in The Anthropic Cosmological Principle, but we’ve seen it already from numerous physicists.

The first anthropic pressure is the probabilistic bias towards chaos, disorder, and inhospitability to life, intelligent life in particular. In most of the multiverse life is impossible. But in some tiny portion, in which we (surprise!) happen to find ourselves, intelligent life just barely was able to evolve out of the muck and acquire enough cognitive complexity to consciously kill each other and compete for mates instead of just doing so mindlessly.

The second anthropic pressure is slightly more speculative. It’s the idea that intelligent species that are too smart wipe themselves out too quickly to really get anywhere. They build self-improving AIs that ignore their creators and tile the cosmic neighborhood with value structures that are a mere shadow of what the programmers originally meant, or launch superintelligent uploads who slowly, and then quickly become obsessed with the idea of constantly stimulating their own pleasure centers to the exclusion of all other pursuits. Both outcomes radically reduce the number of conscious individuals in existence after that point, thereby selecting those quadrants of spacetime out of the anthropic lottery. We’re unlikely to be born into those regions because they are relatively uninhabited, just like we’re unlikely to be born in universes where infant stars have so much gravity that their accretion discs get sucked in before forming stable planets.

We are born in regions that are typical. Industrial civilizations filled with billions of non self-modifying intelligent social animals, apparently. We’re relatively unintelligent because 1) we just evolved from the muck and 2) because we haven’t been clever enough to destroy ourselves yet. Two factors, any one of which alone would be enough to hold the argument up.

But, worry not. There is no reason to despair. These anthropic arguments for our relative stupidity only underscore our potential for growth. We can improve our quality of life to new heights we could never even dream of.

There is an issue of concern, however. If the future is so much more prosperous and populous than today, then why don’t we find ourselves there, instead of here? If out of every 1,000,000 random beings, only one finds itself in civilizations with only a few billion people, then is it just an enormous coincidence that we happen to find ourselves here?

Coincidence is not a satisfactory explanation. There are reasons to believe that this probabilistic issue is a huge problem. It’s called the Doomsday Argument. You can find numerous rebuttals in the Wikipedia article, but many of them are quite subtle, and if you dismiss the argument merely based on its implications, then I think we can justifiably throw out your opinion.

What is your reason for dismissing the Doomsday Argument? Or if you don’t have one, how do you cope?

Dr. Alan Goldstein Interviewed by RU Sirius Tuesday, Apr 24 2007 

Dr. Alan Goldstein explains how the leading artificial life researchers are creating a huge hazard by failing to understand the role of chemistry in molecular engineering and manufacturing, aka, nanotechnology. He says the aliens will not arrive from space, but rather, from our own labs, and that by not understanding this clearly, we put humanity itself at risk.

The Lifeboat Foundation is sponsoring Goldstein’s initiative for raising awareness about nonbiological replicators - the A-Prize. Listen to a separate mp3 interview about the A-Prize here.

Next Page »