Accelerating Future

Humanity is just a tiny spark. If we survive the next few decades and successfully develop superintelligence, that spark will grow into a tremendous fire on the scale of a supernova. If we fall prey to existential risk, that spark will be forever extinguished.

Our present culture, accomplishments, and intelligence are little in comparison to what we have the potential to become by the end of this century. We have the potential to become so radically intelligent that our current selves are mere insects by comparison. This doesn’t mean we should hate or belittle ourselves, just realize that our greatest potential lies in our future as a civilization rather than our present.

The potential for this superlative transformation rests on three simple precepts: that radically greater-than-human intelligences are physically possible, that the speed cognitive activity is dictated by the speed of the underlying processing elements, and that intelligence can exist on a substrate other than proteinaceous neurons. I’ll briefly address these three.

Is greater-than-human intelligence possible? We have no physical evidence for this, because human beings are the only form of intelligence of which we are aware. However, wouldn’t it be incredibly odd if the first species on Earth to display truly general intelligence also happened to possess the highest possible class of that intelligence? Wouldn’t it be odd if genetic engineering, or brain-computer interfacing, and artificial intelligence all contributed nothing to improving intelligence? I’m welcome to hear arguments for why human beings are already in possession of the highest possible intelligence level that the universe permits, if you have any. If you acknowledge that greater-than-human intelligence is possible, then how much greater? Greater than humans as we are over chimps? As we are over mice? As we are over insects? As we are over a thermostat? Discard all anthropocentric pride, and take a real guess.

Is the speed and quality of cognitive activity dictated by the underlying brain? Traditional Christianity, and likely many other world religions (I don’t know enough about them) strongly imply “no”. They (implicitly or explicitly) teach that humans have a God-given level of intelligence and cognitive speed and that manipulating these values by more than a trivial degree is metaphysically forbidden. If these teachings are wrong, it implies that we can ultimately tweak the underlying operation of our neurons, or replace them with faster neurons, and accelerate our thinking speed, perhaps by as much as a dozen orders of magnitude, because our neurons operate relatively slowly. Recently IBM demonstrated a single-atom memory element and molecular switch. If we could port the structure of our biological intelligence to such a device, who could tell how fast we’d then think or what we could do? This is partially a trick question: we’d be able to think faster to a degree proportional to the difference in switching speed between our biological neurons and the new substrate. I.e., pretty damn fast.

Can intelligence exist on substrates other than proteinaceous neurons? As intelligence is clearly a type of information-processing, transforming and integrating sensory information into concepts and outputting actions, there seems to be little reason why not. On a planet where life hypothetically evolves out of, say, methane-based biochemistries, our proteinaceous neurons would be as foreign to them as they to us. All it takes is for there to exist a single example of non-protein-based intelligence in the entire multiverse and we’d know that such entities are hypothetically possible. Why not try building one and see?

Bruce Klein, Outreach Director for the Singularity Institute, recently asked a question as far and wide as he could: “When will AI surpass human-level intelligence?” 50% of the respondents estimated between now and 2050. This post is directed towards that group. If AI is possible before 2050, then that means that intelligence can be ported to a nonbiological substrate, and that quickly, human intelligence will be portable to that substrate as well. That unlocks the possibility of us accelerating our thinking speed, and experiencing much more in much less time. We could experience centuries of time in a single day. This helps us envision how much we have to lose if our species is snuffed out by a technological disaster in the time between now and then.

Of course, if all three of these precepts are false, then we can still do a lot of interesting things in the future, like healing all diseases, wiping out poverty, and extending our lifespans greatly. Flying cars and megastructures. But to me, the most interesting future lies in the possibility of massively boosting our intelligence and thinking speed by transferring our civilization over to new substrates, such as diamondoid nanocomputing.

29 Responses »

(Cross-posted from the Lifeboat Foundation blog.)

A point on human extinction risk analysis.

To look at existential risk rationally requires that we maintain a cool, detached perspective. It’s somewhat hard to think of how this might be done, although watching videos of planetary destruction could actually help! As a detective needs to look at a few crime scenes before he can get experienced and move beyond being a simple gumshoe, existential risk analysts need to view simulations and thought experiments of planetary destruction before they can consider it without flinching. Because it is impossible to acquire experience of human extinction risk, as by definition no one is alive afterwards, we have to settle for simulations.

The reaction of many educated adults to extinction risk discussions reminds me of the reaction kids in my Middle School health classes had to the mention of the word “penis”: adolescent giggling. If I were to get onstage in front of a random audience and start talking about existential risk when they didn’t expect it, using words like “planetary destruction”, they’d probably start giggling, at least in their minds. Obviously, we have a way to mature as a society until we can look calmly at the prospect of our own demise. By resolving to do so yourself, you can be a part of the solution instead of the problem.

Last week a blogger for the Houston Chronicle, Eric Berger, covered my post on immortality and extinction risk, and the immaturity of most of the comments received is expected but also telling. One reader writes that we should hire Will Smith to save the world, another writes: “I don’t worry about this sort of thing, because when it happens, I’ll be dead and won’t care.” Just like how you get to see someone’s true self a little better when they’re a tad tipsy, we get to see what people really think of extinction risk analysis by their anonymous comments on a big website. When people are on the record, they aren’t likely to make pithy comments like those on the blog, but they might be thinking them, and what they say in public is likely to be a dressed-up version of these sentiments. For instance, there’s an article that appeared in The Mercury on the 22nd of April in 2003, “Disastronomer Royal: More Apocalyptic then the Pope”, which exemplifies the reaction to those who take the prospect of extinction risk seriously, referring to Martin Rees in this case. Extinction denialist articles are not hard to find on the Internet: just Google them.

Ideally, existential risk analysis should be getting hundreds of millions of dollars in funding, as the study of global warming does today. Until there are planetary immune systems in place that can respond so quickly and comprehensively that the likelihood of terminal disaster is reduced to practically nothing, existential risk mitigation should be the number one priority of the human species. And the first step is for individuals, such as yourself, to look at the prospect of human extinction in a serious way.

16 Responses »

There are two sides to living as long as possible: developing the technologies to cure aging, such as SENS, and preventing human extinction risk, which threatens everybody. Unfortunately, in the life extensionist community, and the world at large, the balance of attention and support is lopsided in favor of the first side of the coin, while largely ignoring the second. I see people meticulously obsessed with caloric restriction and SENS, but apparently unaware of human extinction risks. There’s the global warming movement, sure, but no efforts to address the bio, nano, and AI risks.

It’s easy to understand why. Life extension therapies are a positive and happy thing, whereas existential risk is a negative and discouraging thing. The affect heuristic causes us to shy away from negative affect, while only focusing on projects with positive affect: life extension. Egocentric biases help magnify the effect, because it’s easier to imagine oneself aging and dying than getting wiped out along with billions of others as a result of a planetary plague, for instance. Attributional biases work against both sides of the immortality coin: because there’s no visible bad guy to fight, people aren’t as juiced up as they would be, about, say, protesting a human being like Bush.

Another element working against the risk side of the coin is the assignment of credit: a research team may be the first to significantly extend human life, in which case, the team and all their supporters get bragging rights. Prevention of existential risks is a bit hazier, consisting of networks of safeguards which all contribute a little bit towards lowering the probability of disaster. Existential risk prevention isn’t likely to be the way it is in the movies, where the hero punches out the mad scientist right before he presses the red button that says “Planet Destroyer”, but because of a cooperative network of individuals working to increase safety in the diverse areas that risks could emerge from: biotech, nanotech, and AI.

Present-day immortalists and transhumanists simply don’t care enough about existential risk. Many of them are at the same stage with regards to ideological progression as most of humanity is against the specter of death: accepting, in denial, dismissive. There are few things less pleasant to contemplate than humanity destroying itself, but it must be done anyhow, because if we slip and fall, there’s no getting up.

The greatest challenge is that the likelihood of disaster per year must be decreased to very low levels — less than 0.001% or something — because otherwise the aggregate probability computed over a series of years will approach 1 at the limit. There are many risks that even distributing ourselves throughout space would do nothing to combat — rogue, space-going AI, replicators that eat asteroids and live off sunlight, agents that pursue reproduction at the exclusion of value structures such as conscious experiences. Space colonization is not our silver bullet, despite what some might think. Relying overmuch on space colonization to combat existential risk may give us a false sense of security.

Yesterday it hit the national news that synthetic life is on its way within 3 to 10 years. To anyone following the field, this comes as zero surprise, but there are many thinkers out there who might not have seen it coming. The Lifeboat Foundation, which has saw this well in advance, set up the A-Prize as an effort to bring development of artificial life out into the open, where it should be, and the A-Prize currently has a grand total of three donors: myself, Sergio Tarrero, and one anonymous donor. This is probably a result of insufficient publicity, though.

Genetically engineered viruses are a risk today. Synthetic life will be a risk in 3-10 years. AI could be a risk in 10 years, or it could be a risk now — we have no idea. The fastest supercomputers are already approximating the computing power of the human brain, but since an airplane is way less complex than a bird, we should assume that less-than-human computing power is sufficient for AI. Nanotechnological replicators, a distinct category of replicator that blurs into synthetic life at the extremes, could be a risk in 5-15 years — again, we don’t know. Better to assume they’re coming sooner, and be safe rather than sorry.

Once you realize that humanity has lived entirely without existential risks (except the tiny probability of asteroid impact) since Homo sapiens evolved over 100,000 years ago, and we’re about to be hit full-force by these new risks in the next 3-15 years, the interval between now and then is practically nothing. Ideally, we’d have 100 or 500 years of advance notice to prepare for these risks, not 3-15. But since 3-15 is all we have, we’d better use it.

If humanity continues to survive, the technologies for radical life extension are sure to be developed, taking into account economic considerations alone. The efforts of Aubrey de Grey and others may hurry it along, saving a few million lives in the process, and that’s great. But if we develop SENS only to destroy ourselves a few years later, it’s worse than useless. It’s better to overinvest in existential risk, encourage cryonics for those whose bodies can’t last until aging is defeated, and address aging once we have a handle on existential risk, which we quite obviously don’t. Remember: there will always be more people paying attention to radical life extension than existential risk, so the former won’t be losing much if you shift your focus to the latter. As fellow blogger Steven says, “You have only a small fraction of the world’s eggs; putting them all in the best available basket will help, not harm, the global egg spreading effort.”

For more on why I think fighting existential risk should be central for any life extensionist, see Immortalist Utilitarianism, written in 2004.

26 Responses »

An Institute for Ethics and Emerging Technologies talk in NYC in May that I missed: Lead Me Not Into Temptation: Folk-Psychological Conceptions of Willpower and Their Implications for Policy. This was at the Human Rights for the 21st Century conference. Vassar offers plausible explanations for the long-standing hostility towards cognitive liberty throughout the world.

Abstract:

“Neither Liberal, Conservative, nor Libertarian political philosophies usually give much explicit attention to the concept of willpower (entirely conceptually seperate from “free will”). However, some examination shows that variation in how it is concieved of appears to be the basis for ideological conflicts between the partisans of different views. Until matters of fact are clarified and resolved, they may appear to be conflicting values, and the apparent conflicts may appear irresolvable. Not only that, the opposing partisans may appear insane. In this presentation I will explain how conceptions of willpowe as abundant, limited, or muscle-like, e.g. limited but renewable and capable of being cultivated and increased, imply different policy proscriptions corresponding to political divides. I will attempt to outline the necessary experiments that should enable us to determine how willpower actually works or to build better metaphors in its place, and will examine the impact of the appearently dominant views with respect to cognitive liberty.”

I consider the talk brilliant, and have praised Michael here before, but it’s really worth listening to this talk closely and understanding what it means. Practically everyone identifies with one of the well-known political alignments. These alignments are ultimately illusions, based on folk psychological theories of society and individuals. Approaching the issues from a more subtle and academically objective perspective (yes, it’s possible) severs the Gordian knot of contemporary political discourse, and brings up the possibility of making actual progress.

Fascinating tidbit: when offered cake in an experiment, those who refused the cake later performing poorly on the Stroop test, involving naming colored words. Michael elaborates on this near 17:00 in his talk.

Again: 99.9% of the people I meet come at political discourse from one of the preexisting camps, and they seem “locked in” - possibly forever, not able to thoroughly analyze the arguments of the opposing side and name why the other side believes these things without creating a straw man. The term “upwinger” has been used to describe someone transcending traditional political polarities. Of course, partisans of the current polarities have a great interest is dismissing even the in-principle existence of such a person, although I think that’s radically wrong. We need to consult cognitive science and evolutionary psychology to look at the underlying causal reasons why people believe the things they do, instead of constantly using superficial explanations “because they’re stupid” that keep us going around and around in perpetual circles of political conflict and idiocy.

Persistent disagreements are irrational. Rational agents will have beliefs that eventually converge. Bayesian wannabes cannot agree to disagree. Aumann’s agreement theorem demonstrates that if people persistently disagree they must not be trying to approximate rationality.

Some of Michael Vassar’s papers are hosted on Accelerating Future at this URL.

14 Responses »

I’m just curious. Here I’m specifically talking about existential risks generally considered to be more than 15 or so years in the future (even though they may in fact be nearer), like self-replicating microbots, recursively self-improving AI, and the like. Do non-immortalists just not look very far ahead, or are they just skeptical that the risks are technologically feasible?

There is somewhat of an overlap between the technologies predicted to lead to radical life extension (nanotech mainly) and the risks themselves, so it would make sense that immortalists are more informed on these technologies, including their risks. But this overlap only goes so far - websites on radical life extension generally address the benefits of medical nanotechnology while largely ignoring the risks. There are also many risks unrelated to life extension: AI, synthetic biology, nanoweapons, etc. So maybe immortalists just care about these risks because they have a much longer expected lifespan and accordingly look further into the future?

The ironic thing is that risks 15+ years away still threaten most of the population today, including all baby boomers. Why is it, on average, that I see more serious attention given to existential risks from the younger (than 40) set than the boomers? (Update: this may be incorrect on my part, based on the bias that my friends tend to be younger, and past involvement with SIAI where many supporters are young. Look at the Lifeboat Foundation donor list and you see many people over 40, plus Martin Rees, Stephen Hawking, Ray Kurzweil, etc.)

Of particular concern are the older Republicans in the United States, pre-boomers, who advocate the development and potential deployment (against Iran) of nuclear weapons. These people have lived the longest with the threat of nuclear war - why is it that they seem the most hawkish about the deployment of nukes? Don’t they understand that a single use of a nuke would set a precedent for further usage by other countries, and that the expected cost of using conventional firepower is much less, even if thousands of soldiers have to die, because it avoids setting a nuclear precedent?

I wish the US would have a policy where it went into future wars with a declaration not to use nuclear weapons, as long as certain conditions are kept (like no one else jumping in). Maybe the world would feel less like it’s being threatened by the US that way, and we could reduce the justifications used by rogue states to acquire the weapons.

18 Responses »

10 Reasons to Live as Long as Possible

1. Because the universe has plenty of room.
2. Because eighty or ninety years isn’t enough to try much out.
3. Because death is so final.
4. Because you’ll get to see what happens next.
5. Because a few hundred friends isn’t enough.
6. Because you can then join the long-term project to make Earth better.
7. Because boredom can’t prevail against new places, ideas, and people.
8. Because aging and death are primitive and inherently unpleasant.
9. Because your loved ones and children don’t deserve to see you perish.
10. Because if you don’t enjoy it, you can end it at any time.

10 Reasons to Enhance Human Intelligence

1. Because stupidity stops being funny pretty fast.
2. Because sitting in a classroom can be torture.
3. Because if we don’t, somebody else will.
4. Because it’s part of our species growing up.
5. Because people are suffering due to their own ignorance.
6. Because people are suffering due to ignorance of the wealthy.
7. Because dumb people don’t know what they’re missing.
8. Because it’s the best way to improve our society.
9. Because our brains were meant to evolve.
10. Because it’s already within our grasp.

10 Reasons to Develop Molecular Nanotechnology

1. Because killing sentient animals is an cruel way to transform grass into meat.
2. Because fossil fuels are a pain in the ass.
3. Because humans can’t fly without strapping ourselves into a rigid hunk of metal.
4. Because our houses and cities could use some embellishment.
5. Because conventional manufacturing destroys the environment.
6. Because people shouldn’t have to perform manual labor if it can be automated.
7. Because biology can do it, so should we.
8. Because everyone deserves food, shelter, and clean water.
9. Because it’s about time we cleaned up the mess we humans have made.
10. Because nanotech will be developed anyway, we might as well develop it right.

10 Reasons to Develop Artificial Intelligence

1. Because human intelligence hasn’t exploited enough possibilities.
2. Because intelligence should be fluid and expandable, not rigid and static.
3. Because we need someone to help us make sense of the data we’re drowning in.
4. Because aliens aren’t showing up, so we should make our own.
5. Because the universe should be infused with intelligence.
6. Because we need new perspectives and thinkers.
7. Because it would be interesting to engineer new emotions.
8. Because sci-fi stereotypes need to be shattered.
9. Because evolution has made us self-deceiving, and we need help to escape the trap.
10. Because AI is coming whether we like it or not, and it’s better to be prepared.

10 Reasons to Learn About Science and Technology

1. Because it has the potential to prevent our extinction.
2. Because so many popular beliefs are empirically unsound.
3. Because it’s the only way to significantly move our entire civilization forward.
4. Because science just means “stuff we know” and technology just “stuff we can do”.
5. Because it’s the foundation and context of all other human affairs.
6. Because it makes the main difference between cavemen and modern man.
7. Because ignorance is nothing to be proud of.
8. Because they’re challenging and useful.
9. Because progress in science and technology is exponentially accelerating.
10. Because our future depends on it.

11 Responses »

Recently, the Lifeboat Foundation’s International Spokesperson, Philippe Van Nedervelde, spoke to the Deputy Editor of Betterhumans, Parish Mozdzierz, on the Lifeboat Foundation, its goals and activities. Here is the first question:

Betterhumans: How did the formation of the Lifeboat Foundation come about?

Philippe Van Nedervelde: Lifeboat Foundation’s founder, Eric Klien, was shaken wide awake by 9/11. The new reality of what we call (exponentially accelerating) “Asymmetric Destructive Capability” (ADC) fully hit him: ever smaller groups of people can create ever more enormous amounts of damage. And all of this thanks to advances in technology. As a bracelet-wearing cryonicist, he knew of the potentials of nanotechnology (having attended MIT Nanotechnology Group meetings in the late 1980s), and that 9/11 was just a taste of things to come. Accordingly, the Lifeboat Foundation was incorporated within months of 9/11.

Read the whole thing here.

If you find the Lifeboat Foundation’s work valuable, please take the plunge and join us!

No Responses »