Accelerating Future

Lately, I’ve been seeing something interesting — valid criticism of the transhumanist project. The concern is decently articulated by the people who are being paid to attack me and other transhumanists, over at The New Atlantis Futurisms blog, funded by the Ethics and Public Policy Center, “dedicated to applying the Judeo-Christian moral tradition to critical issues of public policy”. To quote Charles T. Rubin’s “What is the Good of Transhumanism?”:

While some will use enforcement costs and lack of complete success at enforcing restraint as an argument for removing it altogether, that is an argument that can be judged on its particular merits – even when the risks of enforcement failures are extremely great. The fact that nuclear non-proliferation efforts have not been entirely successful has not yet created a powerful constituency for putting plans for nuclear weapons on the Web, and allowing free sale of the necessary materials. In the event, transhumanists, like “Bioluddites,” want to make distinctions between legitimate and illegitimate uses of “applied reason,” even if as we will see they want to minimize the number of such distinctions because, as we will note later, they see diversity as a good. Of course, those who want to restrict some technological developments likewise look to some notion of the good. This disagreement about goods is the important one, untouched by “Bioluddite” name-calling. The mom-and-apple-pie defense of reason, science and technology one finds in transhumanism is rhetorically useful, within the framework of modern societies which have already bought into this way of looking at the world, to lend a sense of familiarity and necessity to arguments that are designed eventually to lead in very unfamiliar directions. But it is secondary to ideas of what these enterprises are good for, to which we now turn, and ultimately to questions about the foundation on which transhumanist ideas of the good are built.

Yes, diversity is good. But transhumanists have a problem. Diversity is so darn huge, and contains far far more of what would broadly be considered “hideous” than anything beautiful.

I approach the idea of “diversity” from an information theory based perspective. In such a perspective, “diversity” can be achieved by randomly rearranging molecules to achieve a new, unique, “diverse” state. In this view, if absolute freedom to self-modify became possible in a society with sophisticated molecular nanotechnology, then eventually a very large and exotic collective of wireheaded and partially wireheaded beings could emerge. It could be ugly, not beautiful. For a “real-world” example, look at how everyone had great expectations for SecondLife, then it “degenerated” into a haven of porn and nightclubs. While it’s debatable whether a world of porn and nightclubs is a bad thing, it’s obviously not what many in society would want, and I think that an optimal transhumanist future should be appealing to all, not just a few.

Simplistic libertarian transhumanism simply argues, “anything is possible, and everything should be”. Pursued to its logical conclusion, that means that I should be allowed to manufacture a trillion cyborg nematodes filled with botulism toxin and just chill with them. After all, it’s my own choice, what right do you have to infringe upon it? The problem is that that cluster of nematodes would become a weapon of mass destruction if launched into stratospheric air currents for worldwide distribution, and programmed to fall in clusters on major cities where they would inject their toxins into targets which they would navigate to via thermal sensing. My unlimited “freedom” could become your unlimited doom, overnight. The same applies to people in space with the ability to anonymously cloak and accelerate asteroids towards ground targets. Any substantial magnification in human capability raises the same “civil rights” issues.

Many transhumanist writings advocate simplistic libertarian transhumanism. I won’t bother to list any by name, but they’re all around.

A regular commenter here, Sulfur, recently articulated his objection to transhumanism, responding to my recent statement “The latter makes sense, the former doesn’t.”, with regards to solving the flaws of the Homo sapiens default chassis:

The fundamental problem with that sentence is that transhumanists see human body as a problem to solve and they are quick to judge what is needed and what is not. If that would be for them to decide, we already would have done terrible mistakes in augmenting our bodies (”Hell, we don’t need so many genes! let’s get rid of them!” hype-like attitude). Transhumanism uses imperfect tools to perfect human. That can easily lead to disaster. Besides, the most important issue is not weather small changes correcting some flaws are desirable, needed or wanted, but rather to what extend we can change human and not to commit suicide in ambitious yet funny way thanks to augmentation which would radically change our minds, creating new quality.

It’s true — we do see the human body as a problem to solve. After all, the human body can’t even withstand 5 psi overpressure without our eardrums exploding, or intercept rifle bullets without severe tissue damage, which I consider unacceptable. Moving more in a mainstream direction, many transhumanists (a small group of less than 5,000 people with mainstream intellectual influence far beyond their numbers) agree that solving aging is a major priority. After all, Darwinian evolution did not have our best interests in mind when it designed us. As far as I am concerned, the question of whether the human body is a problem to be solved is obvious: it is. The question is not whether or not we need to solve it, but how.

The “how” question is where things can get sticky. Most of human existence is not so crime-free and kosher as life in the United States or Western Europe. Business as usual in many places in the world, including the country of my grandparents, Russia, is deeply defined by organized crime, physical intimidation, and other primate antics. The many wealthy, comfortable transhumanists living in San Francisco, Los Angeles, Austin, Florida, Boston, New York, London, and similar places tend to forget this. The truth is that most of the world is dominated by the radically evil. Increasing our technological capabilities will only magnify that evil many times over.

The answer to this problem lies not in letting every being do whatever they want, which would lead to chaos. There must be regulations and restrictions on enhancement, to coax it along socially beneficial guidelines. This is not the same as advocating socialist politics in the human world. You can be a radical libertarian when it comes to human societies, but advocate “stringent” top-level regulation for a transhumanist world. The reason why is that the space of possibilities opened up by unlimited self-modification of brains and bodies is absolutely huge. Most of these configurations lack value, by any possible definition, even definitions adopted specifically as contrarian positions to try and refute my hypothesis. This space is much larger than we can imagine, and larger than many naive transhumanists choose to imagine. This is especially relevant when it comes to matters of mind, not just the body. Evolution crafted our minds over millions of years to be sane. More than 999,999 out of every 1,000,000 possible modifications to the human mind would be more likely to lead to insanity than improved intelligence or happiness. Transhumanists who don’t understand this need to study the human mind and looming technological possibilities more closely. The human mind is precisely configured, the space of choice is not, and ignorant spontaneous choices will lead to insane outcomes.

The problem with transhumanism is that it has become, in some quarters, merely a proxy for the idea of Progress. Progress is all well and good. The problem is that the idea isn’t indefinitely extensible. The human world is a small floating platform in a sea of darkness — a design space that we haven’t even begun to understand. In most directions lie Monsters, not happiness. Progress within the human regime is one thing, but the posthuman regime is something else entirely. Imagine having First Contact with a quadrillion different alien species simultaneously. That is what we are looking at, with an uncontrolled hard takeoff Singularity. Just one First Contact would be the most significant event in human history, but transhumanists are talking about that times a billion, or a trillion, all at once.

In the comments, Sulfur referenced the “transhumanist mindset which says that upward change is a dogma”. But there is a portion of transhumanists who resist that dogma. Take Nick Bostrom’s “The Future of Human Evolution” paper, very popular among SIAI staff. I believe that Bostrom’s 2004 publication of this paper was a ground-breaking moment for transhumanism, definitive of a schism that has been ongoing since. The schism is between those who see transhumanism as unqualifiedly good and those who see humanity’s self-enhancement as a challenging project that demands close attention and care. Here’s the abstract:

Evolutionary development is sometimes thought of as exhibiting an inexorable trend towards higher, more complex, and normatively worthwhile forms of life. This paper explores some dystopian scenarios where freewheeling evolutionary developments, while continuing to produce complex and intelligent forms of organization, lead to the gradual elimination of all forms of being that we care about. We then consider how such catastrophic outcomes could be avoided and argue that under certain conditions the only possible remedy would be a globally coordinated policy to control human evolution by modifying the fitness function of future intelligent life forms.

I am strongly attracted to the Singularity Institute, Future of Humanity Institute, and Lifeboat Foundation, because I see these three organizations as the cautious side of transhumanism, exemplified by the concerns aired in the above paper. Many other iterations of transhumanism seem to be awkward fusions between SL2 transhumanism and the boilerplate leftist or rightist politics of the Baby Boomer generation. Though even our new President is attempting to engage in post-Boomer politics, the USA Boomer Politics War is so huge that it sucks in practically everything else. It’s pathetic when transhumanists can’t be intellectually strong enough to transcend that. Really, it is a generational war.

As somewhat of a side note, people misunderstand the SIAI position with respect to this question. SIAI seeks not to impose a superintelligent regime on the world, but rather asks, “given that we believe a hard takeoff is likely, what the heck can we do to preserve Human Value, or structures at least continuous with human value?” The question is not easy, and people often misinterpret the probability assessment of a fast transition as a desire for a fast transition. I would desire nothing more than a slow transition. I just don’t think that the transition from Homo sapiens to recursive self-improvement will be very slow. Still, even if it’s fast, value can probably be retained, if we allocate significant resources and attention to specifically doing so.

I believe that there can be a self-enhancement path that everyone can agree on as beneficial. I think there is enough room in the universe to hold diverse values, but not exponentially diverse in the information theory sense. I doubt that intelligent species throughout the multiverse retain their legacy forms as they spread across the cosmos. Inventing and mastering the technologies of self-modification is not optional for intelligent civilizations — it’s a must. The question is what we use them for, and whether we let society degenerate into a mess of a million of shattered fragments in the process.

35 Responses »

How can a small group of people have a big impact on the world? Develop a machine or service that is self-replicating or self-amplifying.

In a mundane way, artifacts such as iPhones and even shovels engage in human-catalyzed self-replication. People see them, then want them, then offer their money for them (or build them themselves, in a few cases), which provides the economic juice necessary to increase production and maintain the infrastructure necessary for that self-replication, like the Apple Store.

Self-replication can be relatively easy as long as the substrate is designed to contain components not much less complex than the finished product. For instance, the self-replicating robot built at Cornell self-replicates not from scratch, but rather from a set of pre-engineered blocks not much simpler than the robot itself. Using a hierarchy of such self-replicators, where each step is relatively simple but results in the creation of more complex components used in the next stage of self-replication, could provide a bootstrappable pathway to self-replicating infrastructures. Such a scheme also makes recycling easier — if a large machine falls apart, perhaps only some of its components need by discarded, and the rest can be reused.

At the root of a substantial number of transhumanists’ wild visions appears to be confidence that self-replicating factories will ultimately be produced. Otherwise, it is hard to imagine how society would acquire the necessary wealth to implement changes of the type that transhumanists discuss. In fact, it appears to me that modern transhumanism evolved in large part out of enthusiasm for the idea of molecular nanotechnology in the mid-1990s. The ongoing philosophical connection of transhumanism to other Enlightenment movements is more of a post hoc project designed to make transhumanism palatable and comprehensible to larger groups.

At its core, I believe that transhumanism’s greatest accomplishment is identifying self-replicating and self-amplifying processes as humanity’s greatest opportunity and hazard of the 21st century — technology with the potential to allow us to transcend our material, physiological, and psychological limitations or, if handled poorly, cause a reprise of the Permian-Triassic extinction. You don’t have to be a transhumanist to appreciate this insight; you only need to be convinced that self-replicating machines are technically plausible at some point in the near or mid-term future. Indeed, a substantial minority of tech-oriented people seem open to the possibility. Here is a poll from a 2005 CNN article on RepRap:

Even more exciting to me than self-replication is the power of self-amplification. I define self-amplification as a growing optimization process that extends its own infrastructure in a diverse way rather than simple self-replication, where “infrastructure” is defined as both core structures and the peripheral structures that support them. Humanity is an interesting edge case here, at the boundary of what I would consider the transition from self-replication to self-amplification. We are able to create diverse artifacts, but our ability to inject diversity into our own bodies and minds through self-transformation or directed evolution is extremely limited.

There is an opportunity here for the development of a mathematical model that quantifies the information and structural content produced by a given self-replicating or self-amplifying entity. Humans like to think that we exhibit nearly infinite variety in the creation of artifacts, but this is untrue. We mostly create artifacts that we have cultural and evolutionary predispositions to create. If we realized how constrained our information-producing tendencies are, it would help us become a more mature species through better self-reflection.

12 Responses »

Yesterday, Good posted the seventh and second-to-last installment of myself and Roko Mijic’s series on the Singularity, “Criticisms of the Singularity”. (My last contribution to the series, “The Benefits of a Successful Singularity”, was promoted to the front page of Digg.) For your benefit, the complete article is reproduced here.

Part seven in a GOOD miniseries on the singularity by Michael Anissimov and Roko Mijic. New posts every Monday from November 16 to January 23.

As was previously discussed in our series, the “singularity” means the creation of smarter-than-human intelligence, or “superintelligence,” a type of intelligence that is impressively more intelligent than humans. Possible methods for its creation include brain-computer interfaces and pure artificial intelligence, among others. Various scientists, futurists, and mathematicians that write about the singularity, such as Ray Kurzweil, Nick Bostrom, and Vernor Vinge, consider such an event plausible sometime between about 2025 and 2050. Among those who consider the singularity plausible, it is widely agreed that the event could alter the world, our civilization, and even our bodies and minds profoundly, through the technologies that superintelligence could create and deploy.

Because the singularity is such a new and speculative idea, and the subject of little academic study, there are people that take practically every imaginable position with respect to it. Some, unfamiliar and shocked by the idea, dismiss it outright or simply react with confusion. Others, such as philosopher Max More, dismiss some of the central propositions after more careful study. A substantial number embrace it openly and without too many qualifications, such as futurist Ray Kurzweil, who seems to expect a positive outcome with a very high probability. My organization, the Singularity Institute, and related thinkers such as philosopher Nick Bostrom, see a positive outcome as possible but not without very careful work towards ensuring that superintelligences retain human-friendly motivations as they grow in intelligence and power.

Criticisms of the singularity generally fall into two camps: feasibility critiques and desirability critiques. The most common feasibility critiques are what I call the Imago Dei objection and the Microsoft Windows objection. Imago Dei refers to Image of God, which is the doctrine that humans are created in God’s image. If humans are really created in the image of God, then we must be sacred beings, and the idea of artificially creating a superior being becomes dubious-sounding. If such a superior being could be possible, then wouldn’t God have created us that way to begin with? Unfortunately for this view, science, experimental psychology, and common sense have revealed that humans possess many intellectual shortcomings, and that some people have more of these shortcomings than others. Human intelligence isn’t perfect as it is; long-term improvements may become possible with new technologies.

The Microsoft Windows objection often surfaces when the topic of superintelligent artificial intelligence is brought up and goes something like this: “How can you be expecting superintelligent robots in this century when programmers can’t even create a decent operating system?” The simple answer is that too many cooks ruin a dish, and operating systems are plagued by a huge number of programmers without any coherent theory that they can really agree on. In other fields, such as optics, aerospace, and physics, scientists and engineers cooperate effectively on multi-million dollar projects because there are empirically supported theories that restrict many of the final product parameters. Artificial intelligence can reach the human level and beyond if it one day has such an organizing theory. At the present time, no such theory exists, though there are pieces that may fit into the puzzle.

Lastly, there are desirability critiques. I am very sympathetic to many of these. If we humans build a more intelligent species, might it replace us? It certainly could, and evolutionary and human history support this possibility strongly. Eventually creating superintelligence seems hard to avoid though. People want to be smarter, and to have smarter machines that do more work for us. Instead of trying to stave off the singularity forever, I think we ought to study it carefully and make purposeful moves in the right direction. If the first superintelligent beings can be constructed such that they retain their empathy for humanity, and wish to preserve that empathy in any future iterations of themselves, we could benefit massively. Poverty and even disease and aging could become things of the past. There is no cosmic force that compels more powerful beings to look down upon weaker beings—rather, this is an emotion that comes from being animals built by natural selection. In the context of much of natural selection it is evolutionarily advantageous to selectively oppress weaker beings, though some humans, such as vegans, have demonstrated that genuine altruism and compassion are possible.

In contrast to Darwinian beings, superintelligence could be engineered for empathy from the ground up. A singularity originating with enhanced human intelligences could select the most compassionate and selfless subjects for radical enhancement first. An advanced artificial intelligence could be built with a deep, stable sense of empathy and even lacking an observer-centered goal system. It would have no special desire to discard its empathy because it would lack the evolutionary programming that causes that desire to surface to begin with. The better you understand evolution and natural selection, the less likely you think it is for Darwinian dynamics to apply to superintelligence.

We should certainly hope that benevolent or human-friendly superintelligence is possible, or human extinction could be the result. Just look at what we’re already doing to the animal kingdom. Yet, by thinking about the issues in advance, we may figure out how to tip the odds in our favor. Human-posthuman synergy and cooperation could become possible.

Michael Anissimov is a futurist and evangelist for friendly artificial intelligence. He writes a Technorati Top 100 Science blog, Accelerating Future. Michael currently serves as Media Director for the Singularity Institute for Artificial Intelligence (SIAI) and is a co-organizer of the annual Singularity Summit.

4 Responses »

From PEA Soup, a blog devoted to policy, ethics, and academia:

Many people teach their small children the myth of Santa Claus: that a magical being who lives at the North Pole brings presents on Christmas Eve. Secondary aspects of the myth are that whether one receives presents is a function of one’s behavior, and that you can communicate with Santa about your preferences. Not only parents, but retail establishments and (I have recently discovered) public schools collude in perpetuating this myth among children of a certain age.

Perpetuating the Santa myth has at least these moral reasons against it:

1. It involves a lot of lying and deception practiced on credulous people.
2. It tends to foster greed in children and contributes to their false impression that one’s happiness is determined by one’s material possessions.
3. In telling children that the quantity and quality of one’s gifts are a function of one’s behavior, when actually they are a function of one’s socio-economic standing and parental temperament, it induces moral complacency in well-off children and false feelings of moral inferiority in less well-off children.

It seems to me that these reasons are sufficient to show that perpetuating the Santa myth is immoral. Most of America strongly disagrees with me on this point. I would be interested to know what the professionals at PEA Soup think.

Lying to children about Santa is just one of the many ways in which parents feel no compunctions about manipulating their children rather than treating them as persons. The only “good” thing about this manipulation is that it is supposedly for the benefit of the child, though that is debatable. One problem with the Santa myth not mentioned above is that the associated manipulation and lies is indicative of a broader pattern of manipulation and lies.

5 Responses »

Here is the commentary. Most of all, I enjoy reviews and comments by outsiders with no contact with our current community. Here are a few quotes and my comments:

It is video conference phone call split screen debate between this Yudkowsky guy who is the head scientist at the Singularity Institute, and Lanier who has been the genius hippy in red dread locks since his early pioneering work with Virtual Reality and artificial vision systems.

Before you click the link, let me frame the debate.

These two guys represent the two extremes of a subtle range of viewpoints on evolution, AI, and human consciousness.

An interesting and subtle range that deserves more popular and academic attention and will get it sooner or later because we are building technologies that produce divisive responses to the relevant philosophical issues.

Jaron’s main criticism of the hard AI camp in this debate is that their strong attachment to finding a way past death and their apriori beliefe in the posibility of resonably building self evolving intelegence together become so rhetorically invasive that they can no longer do objective investigation or engineering… that their beliefs and desires make them “religious”.

Well, Jaron would probably prefer if we didn’t do any objective investigation or engineering, but that’s not true. Remember, as cybernetic totalists, we are totally devoted to our goal. Totally awesome!

From my perspective, Jaron is a nothing more than a (very bright) priest who can’t stop doing science in the basement, and Yudkoswsky is nothing less than a scientist that can’t help wanting to build a God.

Hah! A superintelligence would be like a god. I can vaguely understand why people who don’t regard MNT as plausible would disagree with this, but I never understand why those who do believe that MNT is plausible would.

The fireworks in the video begin at 11:00! I actually agree with many of Jaron’s points in the abstract. I disagree with him when he says that we cannot represent some physical systems in totality or simulate them precisely.

10 Responses »

Robin Hanson found a skeptical Bryan Caplan when the former explained his positions on cryonics to the latter. (”The more I furrowed my brow, the more earnestly he spoke.”) Caplan said:

What disturbed me was when I realized how low he set his threshold for [cryonics] success. Robin didn’t care about biological survival. He didn’t need his brain implanted in a cloned body. He just wanted his neurons preserved well enough to “upload himself” into a computer. To my mind, it was ridiculously easy to prove that “uploading yourself” isn’t life extension. “An upload is merely a simulation. It wouldn’t be you,” I remarked. …

“Suppose we uploaded you while you were still alive. Are you saying that if someone blew your biological head off with a shotgun, you’d still be alive?!” Robin didn’t even blink: “I’d say that I just got smaller.” … I’d like to think that Robin’s an outlier among cryonics advocates, but in my experience, he’s perfectly typical. Fascination with technology crowds out not just philosophy of mind, but common sense.

Hanson responded with an articulate explanation of causal functionalism and the illusory quality of the mind/matter distinction:

Bryan, you are the sum of your parts and their relations. We know where you are and what you are made of; you are in your head, and you are made out of the signals that your brain cells send each other. Humans evolved to think differently about minds versus other stuff, and while that is a useful category of thought, really we can see that minds are made out of the same parts, just arranged differently. Yes, you “feel,” but that just tells you that stuff feels, it doesn’t say you are made of anything besides the stuff you see around and inside you.

Although the argument may seem to be about cryonics on the surface, it is really about the viability of uploading.

6 Responses »

Greg Fish, a science writer with a popular blog who contributes to places like Business Week and Discovery News, has lately been advancing a Searleian criticism of causal functionalism. For instance, here and here. Here is an excerpt from the latter:

A Computer Brain is Still Just Code

In the future, if we model an entire brain in real time on the level of every neuron, every signal, and every burst of the neurotransmitter, we’ll just end up with a very complex visualization controlled by a complex set of routines and subroutines.

These models could help neurosurgeons by mimicking what would happen during novel brain surgery, or provide ideas for neuroscientists, but they’re not going to become alive or self aware since as far as a computer is concerned, they live as millions of lines of code based on a multitude of formulas and rules. The real chemistry that makes our brains work will be locked in our heads, far away from the circuitry trying to reproduce its results.

Now, if we built a new generation of computers using organic components, the simulations we could run could have some very interesting results.

On his blog, he says:

The actual chemical reactions that decide on an action or think through a problem don’t take place and the biological wiring that’s the crucial part of how the whole process takes place isn’t there, just a statistical approximation of it.

This is just another version of vitalism. Computers lack the “vital spark” necessary to create the “soul”, even if they implement the functions of intelligence and self-reflection even more effectively than the biological entity that inspired their creation. But those functions are what create intelligence and self-reflection, not magic chemistry-that-can-never-ever-be-simulated-even-in-principle.

There is quite a bit of fuzziness in chemical reactions themselves, and not all this fuzziness is necessary to implement intelligence or “self-awareness”.

Say we have a molecular dynamics simulation of the brain in complete and utter detail. It behaves exactly the same as the intelligence that it is “simulating”. You can say “it’s just a simulation”, but it can achieve all the same things that the original can, including be your friend or even possibly kill you. In such circumstances, “it’s just a simulation” is quite pointless hairsplitting. Certainly, some atomic configurations are conscious and others are not, but there is no vital force that biological molecules possess that high-resolution simulations of those biological molecules would not also possess.

If it walks like a duck, and quacks like a duck, it’s still possible that it’s not a duck, but if it has a perfect emulation of a duck brain and can walk around in a duck body, then it may as well be a duck.

18 Responses »

CNN has coverage of Dawkinsian atheism. Quote:

“I believe a true understanding of Darwinism is deeply corrosive to religious faith,” Dawkins says in his TED Talk.

For a contrary view, see Shermer. His essay reads more like a document trying to sell evolution to theists than a true discussion of why evolution and theism are deeply compatible.

The Bible says that we were created in God’s image, but evolution shows that the creation of man was in fact based on retention of large amounts of vestigal structure and incidental circumstance. If God designed us, even using evolution, then why did he give us wisdom teeth that crowd our mouths and cause medical complications? Why do our characteristics make good sense in light of an unintelligent, unguided process and absolutely no sense in light of an intelligent, guided process?

4 Responses »

The word “complexity” is a confusing one. There are two types of complexity — the vague, layman’s term, which seems to mean something like a great chain of being (”the more like us humans it is, the more complex it must be”), and the precise, mathematical term, Kolmogorov complexity, which refers to the measure of computational resources needed to specify the object. If you are familiar with the latter concept, that’s what you start to think of whenever someone says “complexity”, and people using the layman’s sense of the term start sounding vague and/or confused. People working in AI tend to mean Kolmogorov complexity when they say “complexity”, so if you hang around with people like that for long enough, it gets ingrained into you.

Since Kolmogorov complexity has a nice mathematical definition, it’s very precise. It turns out that lots of not-so-cool things are really complex, like the structure of bread mold, chaotic fluid eddies, or Hadamard’s billiards. The definition of Kolmogorov complexity is agnostic towards what kind of complexity you mean. A random series of bits a quadrillion digits long is more complex than a human being, but it isn’t particularly more interesting than any other random bitstring. (Update: Actually, I am wrong about this (see the comments), because I was thinking about what I thought of as the “functionally relevant” features of human bodies that make them different from other similar piles of chemicals, but wasn’t considering low-level details like precise atomic configurations. If I do, it’s more like 10²⁹ bits, if we assume that it would take about 143 bits per atom to specify their location, type, electron states, etc. So my revised statement would be “A random series of bits 10³⁰ digits long is more complex than a human being, but it isn’t particularly more interesting than any other random bitstring.”)

Describing, with precise mathematics, what distinguishes interesting-to-human complexity from complexity-in-general is the last great task for our civilization. Once we do that, we’ll have solved AI, and come up with a theory for generating any interesting picture, song, book, or other work of art that we want. We’ll have automated creativity, insight, genius, lateral thinking, and inspiration, all with lifeless algorithms. There will always be an exponential computational universe to explore, so I doubt we’d run out of fun, but we’ll know a lot more about its boundaries and features that we do today, if we choose to have that knowledge.

Humans have a very confined range of interests, algorithmically speaking, so we have a tendency to take complexity-in-general and try to reify it through our reality tunnel into human-understandable complex structures, which often includes faces. Imagine staring into a bright screen with a specific pattern for 10 minutes, then turning to a blank wall. That’s our situation as humans, but the imprinting happened before we were born and is based on subconscious hidden priors. The human species as a whole suffers from the Dunning-Kurger effect in that we think this is normal, even when our hidden priors fail in spectacular and hilarious ways.

The Dunning-Kruger effect is so strong in some humans that they actually believe that their way of thinking is normative, and that no higher qualitative intelligence levels above Homo sapiens exist. This is analogous to thinking that the Earth is the center of the universe.

6 Responses »

A friend and associate of mine, Oxford philosopher Toby Ord, has gained some major coverage on the BBC website. Congratulations, Toby! Toby has pledged 10% of his annual salary, plus any yearly earnings above £20,000, to charities fighting poverty in the developing world. He projects that will amount to about £1M over the course of his career, which he has calculated could save 500,000 years of healthy life.

Toby is participating in what I glibly call “utility war” — a worldwide war not for money or power, but to achieve the greatest good for the greatest number (positive utility). This could be the war to end all wars. A war we can be pleased to fight.

For more information, see Giving What We Can.

No Responses »

Several notes on dreams and the dream-computer interface idea.

Dr J. Allan Hobson, a leading dream researcher with Harvard, is publicizing his hypothesis of dreaming with a press release, “Dreams may have an important physiological function”. According to Dr. Hobson, the function of dreams is physiological — a sort of “mental practice” for the awakened state. Hobson said that “dreams represent a parallel consciousness state that is running continuously, but which is normally suppressed while the person is awake”.

If his hypothesis is correct, it has an important implication for rationality. In the mind sciences, people tend to drastically overweight the significance of the ghost or soul — represented by one’s conscious experience and “free will” in various theories. If progress in cognitive science has taught us anything, is it is that this ghost is both an illusion and far less significant than we, in our vanity, think it is. Hobson’s theory is that dreaming has a physiological function, an unflattering reflection on pet theories that assign dreaming a psychological role. It could also explain why animals without neocortex can dream, because dreaming is not contingent on higher mental functions that are unique to humans. This makes sense, as one would assume there to exist far more evolutionary complexity dependent on and associated with evolutionarily ancient features rather than extremely recent evolutionary features like general intelligence, though there are a surprising number of features seemingly uniquely associated with general intelligence.

Dr. Hobson is also known for Dream Debate, a 154-minute DVD video where he admirably pounds nails into Freud’s coffin, taking on the dream theory that just won’t die. Freudian dream interpretation is a example of apophenia, where instead of distrusting yourself and demanding empirical evidence for your theory, you just make shit up that sounds cool and pass it on to others. Apophenia is appealing because anyone can do it, and accordingly, it underlies the delusions of psychic phenomena, “synchronicity”, believing God is talking to you when you pray to him, the Forer effect, which underlies astrology and fortune-telling, and — let’s not forget — Freudian psychology.

Another dream-related update has to do with better brain-computer interfacing technologies being worked on by Ed Boyden at the MIT Media Lab. When I wrote “Brain-Computer Interfaces for Manipulating Dreams”, I assumed that microelectrode arrays would be used to interface with the brain on a finely detailed level, accompanied by tiny holes drilled into the skull which would be carefully resealed. Of course, some of my readers recoiled in horror at this idea, even if it would allow them to become oneironauts of a most magnificent variety. However, thanks to Dr. Boyden’s new approach of stimulating neurons using optics, we’d still have to carve holes in the skulls (to get the optics in), but they would interface optically rather than electrically, which is significantly less invasive. I notice that Wired coverage of Boyden earlier this year is very brief. No one seems to have jumped on this story yet, except for Technology Review, which is pretty amazing.

High-throughput circuit screening of intact brains, which Boyden’s work could enable a cognitive science revolution of the highest order. Here it is, creeping up on us, and no journalists are giving it the coverage it deserves… there are plenty of other stories like this, but unfortunately I don’t have the time to write them all up because there aren’t enough hours in the day.

As always, happy dreaming!

One Response »

To follow up on the previous post, I think that the critique by Massimo Pigliucci (a philosopher at the City University of New York) of David Chalmers’ Singularity talk does have some good points, but I found his ad hominem arguments so repulsive that it was difficult to bring myself to read past the beginning. I would have the same reaction to a pro-Singularity piece with the same level of introductory ad hominem. (Recall that when I was going after Jacob Albert and Maxwell Barbakow for their ignorant article on the Singularity Summit, I was focusing on their admission of not understanding any of the talks and using that as a negative indicator of their intelligence and knowledge, not insulting their hair-cuts.) If anything, put the ad hominem arguments at the end, so that they don’t bias people before they’ve read the real objections.

Pigliucci is convinced that Chalmers is a dualist, which is not exactly true — he is a monist with respect to consciousness rather than spacetime and matter. I used to be on Dennett’s side of the argument and believed there was no hard problem to speak of, but eventually I was moved to somewhere in-between Chalmers and Dennett, and really do believe that there is an interesting hard problem to be solved, but I doubt that solving it will require the introduction of new laws of physics or ontological primitives. I understand why there are people skeptical of the relevance of Chalmers’ theories of consciousness, but the ideas are quite subtle and it took me 2-3 reads of his landmark paper before I started to even pick up on the concept he was trying to transmit. It may be that Pigliucci does understand Chalmers’ ideas and considers them useless anyway.

Moving on to the actual critique, Pigliucci accuses Chalmers of saying that because computers are getting faster, we can extrapolate that to say that means that AI will eventually happen. I think I do vaguely agree with Chalmers on that one, though the extrapolation is quite fuzzy. Since brains are machines that behave according to (as yet unknown) principles but known basic laws (physics and chemistry), faster computers would surely facilitate its emulation, or at the very least the instantiation of its basic operating principles in another substrate. I’m not sure why this is controversial, unless people are conceiving of the brain as including a magical-sauce that cannot be emulated in another finite state machine.

Even if we don’t yet understand intelligence, as Pigliucci points out, that doesn’t mean that it will remain unknown indefinitely. Chalmers even points out in his talk that he thinks it will take hundreds of years to solve AI. My view is that if anyone confidently says that AI will very likely not be possible in the next 500 years, they’re being overconfident and likely engaging in mystical mind-worship and a desire to preserve the mystery of the mind due to irrational sentimentality. Given the scientific knowledge we’ve gained over the last 500 years (practically all of it), it’s quite far-fetched to say confidently that intelligence will elude reverse-engineering over the next 500 or so years. If biology can be reverse-engineered on many levels, so will intelligence.

Pigliucci then points out that Chalmers is lax on his definitions of the terms “AI”, “AI+”, and “AI++”, which I agree with. He could use at least a couple more slides to define those terms better. Pigliucci then argues that the burden of proof of the points that Chalmers argues for is on him because he has an unusual claim. I agree with that also. Chalmers is approaching an issue as philosophy when what it really could use are detailed scientific arguments to back it up. On the other hand, within groups where these arguments are already accepted (like Singularity Summit), philosophy is indeed possible. Some philosophizing has to rest on scientifically argued foundations that are not shared in common among all thinkers. Isn’t it exciting how philosophy and science are so interdependent and how one can just perish without the other?

I disagree with Pigliucci that the “absent defeaters” points are not meaningful. Chalmers is obviously arguing that something extraordinary would need to happen for his outlined scenario not to occur, and that business as usual over the longer term will involve AI++, rather than its absence. “Defeaters” include things like thermonuclear war, runaway global warming, etc., which Chalmers did concretely point out in his talk (at least at the Singularity Summit version). Pigliucci says, “But if that is the case, and if we are not provided with a classification and analysis of such defeaters, then the entire argument amounts to “X is true (unless something proves X not to be true).” Not that impressive.” Maybe Chalmers should have spent more time describing the defeaters, but I don’t think that all arguments of the form “X is true (unless something proves X not to be true)” are meaningless. For instance, in physics, objects fall at 9.8 m/s² unless there is air friction, unless they get hit by another object in mid-fall, unless they spontaneously explode, etc., and the basic law still has meaning, because it applies enough to be useful.

I agree with Tim Tyler in the comments that defining intelligence is not the huge issue that Pigliucci makes it out to be. I do think that g is good enough of an approximate definition (is Pigliucci familiar with the literature on g, such as Gottfredson?), and asking for unreasonably detailed definitions of intelligence even though everyone has a perfectly good intuitive definition of what it means seems to just be a way of discouraging any intelligent conversation on the topic whatsoever. If one would like better definitions of intelligence, I would strongly recommend Shane Legg’s PhD thesis Machine Superintelligence, which gives a definition and an good survey of past attempts at a definition during the first part. I doubt that many will read it though, because people like it when intelligence is mysterious. Mysterious things seem cooler.

Pigliucci then says that AI has barely made any progress over the last few decades because human intelligence is “non-algorithmic”. You mean that it doesn’t follow a procedure to turn data into knowledge and outputs? I don’t see how that could be the case. Many features of human intelligence have already been duplicated in AIs, but as soon as something is duplicated (like master chess), it suddenly loses status as an indicator of intelligence. By moving the goal posts, AI can keep constantly “failing” until the day before the Singularity. Even a Turing Test-passing AI would not be considered intelligent by many people because I’m sure they would find some obscure reason.

Pigliuci continues:

After the deployment of the above mentioned highly questionable “argument,” things just got bizarre in Chalmers’ talk. He rapidly proceeded to tell us that A++ will happen by simulated evolution in a virtual environment — thereby making a blurred and confused mix out of different notions such as natural selection, artificial selection, physical evolution and virtual evolution.

I agree… sort of. When I was sitting in the audience at Singularity Summit and Chalmers started to talk about virtual evolution, I immediately realized that Chalmers had not likely studied Darwinian population genetics, and was using the word “evolution” in the hand-wavey layman’s sense of the word rather than the strict biological definition. If I recall correctly, someone (I think it was Eliezer) got up at the end of Chalmers’ talk and pointed out that creating intelligence via evolution would require a practically unimaginable amount of computing power, simulating the entire history of the Earth. Yet, I don’t understand why Pigliucci believes that such a thing would be impossible in principle — if evolution could create intelligence out of real atoms on Earth, then simulated evolution could (eventually, given enough computing power) create intelligence out of simulated atoms. Of course, the amount of computing power required could be prohibitively massive, but to argue that reality cannot be simulated precisely enough to reproduce phenomenon X just means that we either don’t know enough about the phenomenon to simulate it yet, or we lack the computing power, not that it is impossible in principle. Science will eventually uncover the underlying rules of everything that it is theoretically possible to uncover the rules of (for instance, not casually disconnected universes), and that includes intelligence, creativity, imagination, humor, dreaming, etc.

Pigliucci then remarks:

Which naturally raised the question of how do we control the Singularity and stop “them” from pushing us into extinction. Chalmers’ preferred solution is either to prevent the “leaking” of AI++ into our world, or to select for moral values during the (virtual) evolutionary process. Silly me, I thought that the easiest way to stop the threat of AI++ would be to simply unplug the machines running the alleged virtual world and be done with them. (Incidentally, what does it mean for a virtual intelligence to exist? How does it “leak” into our world? Like a Star Trek hologram gone nuts?)

The burden really is on Chalmers here to explain himself. “Leaking out” would consist of an AI building real-world robotics or servants to serve as its eyes, ears, arms, and legs. Pigliucci probably thinks of the virtual and physical worlds as quite distinct, whereas someone of my generation, who grew up witnessing the intimate connection between the real world and the Wired views them more as overlapping magisteria. Still, I can understand the skepticism about the “leaking out” point, and it requires more explanation. Massimo, the reason why unplugging would not be so simple is that an AI would probably exist as an entity distributed across many information networks, yet that is my opinion, not Chalmers’. From Chalmers point of view, I think he might be concerned that the AI would simply deceive the programmers into believing that it was friendly, therefore long-term evaluations in virtual worlds are necessary. Therefore, the unplugging would not be that simple because we wouldn’t want to unplug the AI, because we could be deceived by it.

Pigliucci says:

Then the level of unsubstantiated absurdity escalated even faster: perhaps we are in fact one example of virtual intelligence, said Chalmers, and our Creator may be getting ready to turn us off because we may be about to leak out into his/her/its world. But if not, then we might want to think about how to integrate ourselves into AI++, which naturally could be done by “uploading” our neural structure (Chalmers’ recommendation is one neuron at a time) into the virtual intelligence — again, whatever that might mean.

Massimo, he is referring to the simulation argument and the Moravec transfer concepts. The simulation argument can be explored at simulation-argument.com, and the Moravec transfer is summarized at the Mind Uploading home page. I know that these are somewhat unusual concepts that should not be referred to so cavalierly, but you might consider reserving your judgment just a little bit longer until you read academic papers on these ideas. Mind uploading/whole brain emulation has been analyzed in detail by a report from the Future of Humanity Institute at Oxford University.

Pigliucci starts to wrap up:

Finally, Chalmers — evidently troubled by his own mortality (well, who isn’t?) — expressed the hope that A++ will have the technology (and interest, I assume) to reverse engineer his brain, perhaps out of a collection of scans, books, and videos of him, and bring him back to life. You see, he doesn’t think he will live long enough to actually see the Singularity happen. And that’s the only part of the talk on which we actually agreed.

Yes, it makes sense that we’d reach out to the possibility of smarter-than-human intelligences to help us solve the engineering problem of aging. Since human biochemistry is non-magical (just like the brain — surprise!) it will only be a matter of time before we start figuring out how to repair metabolic damage faster than it builds up. I’m quite skeptical about Chalmers being genuinely revived from his books and talks, but perhaps an interesting simulacra could be fashioned. While we’re at it, we can bring back Abe Lincoln and his iconic stovepipe hat.

Pigliucci’s conclusion:

The reason I went on for so long about Chalmers’ abysmal performance is because this is precisely the sort of thing that gives philosophy a bad name. It is nice to see philosophers taking a serious interest in science and bringing their discipline’s tools and perspectives to the high table of important social debates about the future of technology. But the attempt becomes a not particularly funny joke when a well known philosopher starts out by deploying a really bad argument and ends up sounding more cuckoo than trekkie fans at their annual convention. Now, if you will excuse me I’ll go back to the next episode of Battlestar Galactica, where you can find all the basic ideas discussed by Chalmers presented in an immensely more entertaining manner than his talk.

I disagree that the topics investigated by Chalmers — human-level artificial intelligence, artificial superintelligence, safety issues around AI, methods of creating AI, the simulation argument, whole brain emulation, and the like — are intellectually disrespectable. In fact, there are hundreds of academics who have published very interesting books and papers on these important topics. Still, I think Chalmers could have done a better job of explaining himself, and assumed too much esoteric knowledge in his audience. A talk suited to Singularity Summit should not be so casually repeated to other groups. Yet, it’s his career, so if he wants to take risks like that, he may have to pay the price — criticism from folks like Pigliucci, some of whose gripes may be legitimate. I also think that Pigliucci probably speaks for many others in his critiques, which is a big part of why I think they’re worth taking apart and analyzing.

2 Responses »