No Heaven (Except What We Make)
It's easy if you try
No hell below us
Above us, only sky
- Lennon
Trigger warning: If you're a religious person, you may perceive what I'm about to write as an affront. It's for you to decide whether to keep reading, or skip to something else. There is plenty of other material out there to focus on.
I’ve debated whether I should put this material out there, as some of it is inherently depressing and/or scary to contemplate. They’re not thoughts unique to me, though (I lay claim to no special insights), and thinking these matters through as a species may be therapeutic as well as necessary. And it’s not all bad. This post has been brewing in my head for some time. It could be emerging now as an extremely introspective person's version of a midlife crisis. I did turn 45 in February, after all.
I've held a variety of beliefs over those years. Growing up, I believed there was an all-knowing, all-seeing God who could directly perceive my thoughts. Later on I was convinced that gray aliens were abducting people. At another time I believed in reincarnation, that our souls survive death and are granted new life on Earth. I believed that prayer had real power to change the world, and traumatic world events generated psychic wavefronts that could be measured electrically. I believed that near-death experiences were proof of the soul's existence. I hold none of those beliefs today. Some of them are embarrassing to own up to.
Much earlier, I believed fervently in the existence of Santa Claus. Most people can relate to this; belief in Santa as a kid is something we can collectively chuckle about and shrug off. Later we learn that belief in Santa is a proxy for something larger, like belief in God. Santa is God for children, rewarding with presents and punishing with coal, rather than dishing out eternal heaven and hell. We get older, and we put away childish things.
But Santa Claus falls into a special category, a belief that's no longer socially acceptable to hold by the time you reach adulthood. Most of us cling to other supernatural beliefs, often for our entire lives. Nobody's told that belief in God is something they need to outgrow.
By contrast, my ground state appears to be agnosticism. I've glommed onto this belief system or that one, but in time I drift back into neutral. How can we know what's “out there”?
Lately, I've been taking the atheist point of view more seriously. Maybe there isn't anything “out there”.
Those who hold a particular religious belief strongly, be it Christianity or Judaism or Islam, or Buddhism or Hinduism, or anything else -- they all think that their own belief system is true and correct. All others seem to them... at best, a little off, odd, alien or ridiculous, at worst offensive and blasphemous. Given how many religions exist, that amounts to most people discounting the beliefs of most other people. For someone like me, who tends to float along in agnosticism, that makes all the religions look slightly silly.
I read Michael Shermer's book The Believing Brain and thought he made some excellent points. Our brains evolved for the same reason everything else about us did: to perpetuate the species. That's it. Evolution, being a purely mechanical process, doesn't care about anything -- not our happiness, not our ability to seek truth, not even our individual longevity except insofar as it aids in reproduction. Evolution has no goals. We are here because our ancestors had a good set of survival tools, and the brain just happens to be a very compelling tool.
How does our brain keep us alive? By noticing patterns. But importantly, the cost of detecting a false pattern is relatively low, while the cost of missing a real pattern can be high. Better to startle needlessly at the wind in the bushes, than to discount the rustle in the grass caused by an approaching predator. So we see and hear patterns everywhere. We have an innate tendency to believe things easily (this follows that), and let them go reluctantly. Our brains find patterns even in purely random noise: faces in the clouds, murmurs in the wind, reversed song lyrics. Randomness makes fools of us all. Presented with truly unpredictable behavior, we'll detect false patterns and swear we've cracked the mystery. Gamblers imagine they see patterns in the outcomes of slot machines, or that they win more often in their lucky underwear. Mind you, we don't believe in luck once we have a solid solution to a problem. Superstition is a crutch we employ whenever we have insufficient knowledge or insufficient control.
Once we see a pattern, it's reinforced through a feedback loop. We're riddled with biases, such as confirmation bias: we're adept at detecting tiny shreds of evidence that bolster our existing beliefs, while discounting large and even obvious piles of evidence to the contrary. We suffer from loss aversion: we'll work harder to keep from losing a dollar than we will to gain a dollar. And for the sake of group cohesion (another handy trick for keeping a tribe alive in hostile conditions), we also tend to believe authority figures.
Psychology may be one of the most important fields today. We're starting to understand how we think, how we make decisions, and why we so often make poor ones. Making better decisions is basically the definition of wisdom, and it's hard to imagine a more worthwhile investment.
The most difficult beliefs to reconsider dispassionately are those central to our identity. Many people do feel this way about their religions. If I believe in the Flying Spaghetti Monster, I'm likely to describe myself as a Flying-Spaghetti-Monsterist. (Edit: Actually they prefer "Pastafarian".) A perceived attack on the idea can feel like an attack on my person, inspiring a fight-or-flight response little different than a pointed gun would inspire. Thus, I will naturally dig in my heels, reach deeper for smaller scraps of evidence in my belief's favor, work harder to find the slightest flaws in the counter-evidence so I can justify rejecting it.
Religion has other things going for it. The idea of death is inherently terrifying when we consider it applying to ourselves, laden with grief when it applies to loved ones, tragic when it applies to everyone. And ultimately it does apply to everyone. How tempting to think that death is not truly the end. That's a reason to want it to be true, but not a reason it might actually be true.
Shermer posits that we are "natural-born immortalists", that our brains are built such that they tend to believe in an afterlife. The argument's a bit complicated to explain in full, but includes our ability to construct a "sense of self" apart from our body, and then to think of that self, that "mind", as something that could exist without the body. We can imagine ourselves in other places, and doing other things. We can also imagine events past our own demise, which lets us do things like buy life insurance for our families. That is evolutionarily useful. Imagining what it would be like to still exist after shedding the body like a skin -- perhaps less so. But the ability to imagine it is not evidence that a mind can exist without a brain to do the thinking. Maybe it just feels that way.
You know what’s hard to imagine? Not existing at all. I’ve been trying lately, to picture what it would be like not to exist or think or feel, and of course I can’t; there is nothing to picture in that picture. There’s no evolutionary advantage to it either, to being able to imagine nothingness. I’m not built for it, so it’s literally unimaginable. A failure of imagination makes the possibility seem unlikely, because imagining things is how we weigh possible futures. Atheists have a hard sell.
Do we have evidence of life after death? Often, we think we do -- religions start from such seeds of evidence. But we now recognize our own tendency to detect false patterns. Couple that with wish fulfillment and imagination failure, and we have strong reasons to suspect that religious beliefs have no substance behind them. Our brains detected false patterns, ages ago, and we latched onto them like shipwreck survivors clinging to flotsam in an ocean, still clinging centuries later. We even liked the ideas enough to fight wars over their particulars. My piece of floating junk versus yours.
With everything else in life, science has been remarkably good at sifting truth out of chaos. It takes time, individual scientists make mistakes, but the rules transcend individuals and their egos. Science insists on the measurable and the falsifiable: if you can't come up with evidence either for or against, then it's irrelevant. And science has been most effective when applying Occam's Razor: all else being equal, favor the simpler explanation. It's simpler to believe the Universe has no opinionated God running the show. There is no way to prove there isn't one... but the burden of proof is on those claiming He exists.
Furthermore, there is an overarching pattern to scientific progress that is hard to miss. The arrow points always in one direction: We Are Not Special. First we believed the Earth was the center and the Sun went around us. Then we learned the Sun was the center. Then we learned the Sun is just one of many stars in the galaxy... and just one of a staggeringly large number of stars scattered around the vast universe. Recent guesses put the number as high as a sextillion -- that’s 1,000,000,000,000,000,000,000 -- and recent exoplanet research suggests there may be, on average, a planet for every star. We learned that here on this planet, this “pale blue dot”, we're not the apex of evolution; we're just pretty darn successful apes with hair loss. Challenging our sense of unity, we learned we are made up of many cells, each a small living creature... and then we learned we also depend for our health on bacteria that don't even share our DNA. We learned that as a species, we've only existed for a blip of time relative to all creation. We learned that other species on this planet can do almost everything we can do. We learned that we are a fragile part of a fragile ecosystem, that we could push far enough out of alignment with our technology to drive ourselves extinct unless we're careful.
It follows, then, that we also wouldn't be special in the sense of having immortal souls. Given that there’s no firm dividing line between us and other animals (we can find apes similar to us, other animals similar to them, etc.), the idea of a soul makes little sense unless all animals have them -- but then what about plants, what about bacteria, what about viruses? Does any of that make sense? Or do we invoke Occam’s Razor again? Do we not require that hypothesis?
Since people continue to cling to supernatural beliefs, we keep searching for evidence in support of them. Sometimes we think we've found it -- hey, look at this ghost video I recorded. But the tendency there is for subsequent experiments to institute tighter controls, weeding out false positives. If the phenomenon were real, results would stubbornly converge on a nonzero result. Instead, they get closer to zero when more controls are added. This suggests that the noise, the false positives, are the only evidence in favor, and that the phenomenon isn't real.
There will always be particular examples we can’t otherwise explain. Hey, you haven’t come up with a better explanation for my ghost video! But that just leaves you with an interesting anecdote. The burden of proof, again, is on those making the claim, and an anecdote isn’t proof. You need lots of proof, repeatable proof, and also a viable explanation. Over and over, attempts to do just that, to collect up all the anecdotes and form a pile of real evidence, have met with failure: the pile melts when you look hard at it. The only reason we keep trying is because we want so badly for it to be true. (Or because we’re trying to sell books, or make money as a spoon-bender, or get you to call a 900 line for advice on your future, or…) A lot of scientists wish we’d stop wasting our effort this way, and redirect it toward something more useful.
I watched a YouTube video recently by the excellent CGP Grey, about the Trouble with Transporters (as depicted in Star Trek). In passing, he mentioned the possibility of people having souls, a dilemma for anyone being taken apart in one place and put back together in another. But he noted the purported soul's impact on reality is unmeasurable. A thing which has no measurable impact on reality could still “exist”, but it wouldn't matter. So it's not worth our time to think about.
OK, perhaps the spiritual world exists and has no impact on this reality, but does affect what we experience after death. Heaven or Hell, perhaps. What then? In that case, we could know nothing about what that next realm is like. The endless debate, this piece of flotsam or that, is ultimately pointless. Most likely, every religion ever invented is wrong, because the number of possibilities is infinite. We have no way of knowing what impact, if any, our actions in this life might have on the next. This, in my opinion, is the biggest problem with Pascal's Wager. You could assume it's safer to believe in God "just in case" He exists, so you can enjoy an eternity of Heaven. But God might turn out to be a nasty character more akin to Cthulhu, leaving you better off not believing in Him lest you draw His attention. Or He might deem you irrelevant. There could be a trillion gods. There could be just a massive soup of collective consciousness. There could be something utterly inconceivable to us here. Or again, there could be nothing. That option is the only one that gets special consideration thanks to Occam’s Razor.
We can't know what's coming without evidence here and now. The harder we look for evidence here and now, the more it seems like there isn't any. We may as well assume there is no next life, that one is all we get.
Seriously contemplating an existence in a purely physical universe, with no spiritual component externally offering meaning, and no promise of life after death, hasn't gotten less terrifying the more I've thought about it. I think it's made me a more cautious driver, though.
Maybe the idea needn’t be so scary or sad. My flailing attempts to picture oblivion fill me with a sort of yawning abyss feeling, like the ground has opened under me. But not-existing wouldn’t feel like that, or like anything. So, “there is nothing to fear but fear itself”. I’m only afraid of death if I choose to think about it as scary. In a way, this is freeing. I don’t have to worry about how my actions in life might be judged afterward. I can focus on living the life I have, and I can live it to please myself and those I care about. (My loved ones, my friends, and more diffusely, every person out there -- they are all like me in basic ways.) What I’m left with is a milder wistful feeling, of wishing there could be more. An afterlife would be nice, the chance to reflect, possibly to watch what else unfolds; and certainly reincarnation would be nice too, a chance to take the ride over again. If I can’t have those things, then simply a longer life is what I’d wish for. Albert Einstein said one life was enough for him, but of course he achieved a lot.
Does it matter if that belief in an afterlife is false? There might be advantages to believing something that's false, if it comforts us, if it leads us to act in ways that serve us. Indeed, we know this is the case. Optimists tend to be happier than pessimists, and more healthy. This utilitarian suspension of disbelief in life after death might fall into that category.
This is a tricky proposition for me, especially now, with all that's been going on in our politics. The supposedly greatest nation on Earth has elected a con man, compulsive liar, demagogue, raving narcissist, and bigot to its highest office. An unapologetic gaslighter, he has made a habit of denouncing respectable news outlets and contradicting well-established facts. In this poisonous atmosphere, I want more than ever to honor truth and carefully weigh evidence.
There’s also the flip side: A belief in an afterlife, whatever your beliefs on how to get there, diminishes the importance of what happens in this one. Religions provide comfort this way, but also engender passivity. You can accept maltreatment and unfairness because you believe it’ll be rectified in the next life. Rejecting that promise of posthumous justice may be fuel for the fire, driving people to stand up for themselves and demand fair treatment and better solutions in this life.
What's more, religious beliefs impact on moral questions. Almost all religions have strong opinions on how its adherents should act. Often those behaviors are positive, but not always. Many religions have self-serving rules, enriching the priesthood, putting them beyond reproach, giving them power which they abuse. At times, religious beliefs have led well-meaning people to torture nonbelievers, even kill them, on the theory that they were saving immortal souls. And of course, the rules in one religion frequently contradict another's. Sometimes the rules are internally inconsistent, or incoherent. Sometimes, more importantly, they conflict with new information we've learned thanks to science. Better to have our eyes open, and be free of imaginary constraints.
Religion frequently presents a simplistic moral view: this act is holy, that act is sinful. This person is good, that one is bad. It's not unlike the caricatures of evil villains in old movies, the idea of a person who does nasty stuff because they enjoy it, or because Satan is making them do it. Rather, we're all humans with similar brains. There are variations, such as our perception of pain: what one person can bear, another cannot. Some people are more creative, more hard-nosed, more compassionate, more pragmatic, more optimistic, more ambitious, more patient. Some variations are harmful, leading to divergent behaviors which we term mental illness. But if you believe there's no spiritual puppet master behind the scenes, the body (brain included) must be all there is to the person. If you could correct the brain, through medicine and therapy, the person's behavior would be socially acceptable again. No concept of evil or sin is required.
In this model, the idea of free will becomes suspect too. Our decision-making is determined by exactly who we are at a given moment, and the environment we're reacting to. Who we are is a combination of our genes, and the experiences that shaped us. Supposing there is no non-material plane, there can be no outside force that would allow us to make decisions "freely". We still make decisions, of course. We choose to eat spaghetti instead of asparagus, say, because we prefer it, but we don't choose our preferences. I write this post because I want to, but I don't choose to want to. I am a product of everything that came before me, and everything around me, and the processes inside of me which I cannot directly control, and of randomness. I’m a very complicated product of all that, one that’s hard or impossible to predict, but that doesn’t mean I can violate the natural order in the course of making a decision. Every choice I make has to be consistent with the rest of physics.
Note that “free will” should not be conflated with “willpower”. The latter must logically be a product of the brain. We’re not entirely sure how to define it, much less where to find its source. But the ability to stand firmly behind decisions in spite of setbacks need not be a mystical quality. It can derive from the structure of the brain, along with every other aspect of one’s personality.
Setting aside the cacophony of competing religions, and focusing on the implications of a purely physical universe, leads to a moral outlook based on psychology. We know that people tend to be more generous when we have enough to be comfortable. We make poor decisions in times of scarcity, because our brains fixate on what we lack. We'll break rules when it's necessary to survive. We respond to incentives, distorted by the fun-house prism of all those biases. We've learned from the Milgram experiment and the Stanford prison experiment that otherwise "normal" people's behavior can become disturbing in certain circumstances. We know adults will often emulate abusive behaviors if abused as children. Pretending all these psychological realities can be countered by an invisible "free will" would be fruitless, and more than that, it would be cruel. If we want to encourage certain behavior, we have to engineer the circumstances that lead to that behavior. That means societal fixes or medical ones (pharmaceuticals or therapy), not righteous upbraiding.
For example, one might propose that a drug addict ought to choose to stop doing drugs, because doing drugs is immoral. If that's the end of the proposal, you haven't solved the problem. Throwing addicts in jail just causes additional suffering to them and their families, and wastes taxpayer dollars. Instead, seek out the reasons why a person is addicted, and find things to change in his/her life that could help. Safe injection sites may be part of the solution. If you don't like that idea, let's hear a better one that works, or that we can at least try. Don't waste everyone's time repeating past mistakes.
This, in turn, calls into question the very concept of "justice", the idea that it's right to punish someone for how they've acted. That too is based on the idea that the person freely chose to misbehave. It would be better to view the person as a malfunctioning machine (for the body and its brain are just imperfect physical machines), and attempt to find and correct the damage that led to the undesirable behavior. Maybe that damage is internal, maybe it’s environmental. One cause we can rule out is a defect of the soul.
Gods are not only respected and feared, but also deeply loved. It’s little wonder, then, that atheists are considered one of the most widely disliked religious minorities in America. It’s as if they are trying to take everyone else’s Santa Claus away. Atheists are also widely distrusted. Discounting God is horrifying to a religious person partly because they usually believe their morality is derived from their religion. Without that anchor, surely, we could choose nearly any moral framework conceivable. Who's to say we wouldn't adopt a framework that admits of all sorts of awful behavior?
Here too, psychology has some things to say. We evolved as a social species, and certain moral rules are apparently baked into us. We're very good at detecting cheaters and freeloaders, for example -- and that sense of justice, unlike the abstract one from earlier, is pretty simple. We expect people to pull their weight, and generosity by others engenders gratitude and inspires reciprocation. In game theory terms, we survived in part because we got together in tribes and cooperated. Today we employ new mechanisms, such as money and politics, to allow us to cooperate on far grander scales, achieving trips to the moon and other such endeavors that would be otherwise impossible.
Living in a material world does not have to mean the end of love, or courage, or beauty. Scientists are fond of pointing out that knowing, for example, how romantic attraction develops in the brain does not diminish the experience of it. Knowing that hormones play a crucial role in puberty doesn’t stop that from happening either. It just gives us a better shot at understanding the phenomenon, so we can have the best possible outcome when raising teenagers.
Still, there are real problems here. Our tribal behavior makes us treat people who are different as "The Other". We discriminate based on skin color, and it's not always done consciously. We treat men and women differently, in totally irrational ways. We dehumanize people who think differently from ourselves, or who simply live on the wrong side of an imaginary line. Human nature is deeply flawed in this respect, but frankly, religion has done little to help. All too frequently, it's just one more category to discriminate against. My piece of floating junk or yours.
Here, I'd like to reference Steven Pinker's Enlightenment Now, and remind the reader of the overarching idea that rational thinking has done a lot for us. Read the book if you don’t believe it. (Read it if you do believe it. It’s a great book.) We have come a long way and can continue to progress. We just have to be mindful, and humble, and keep correcting our mistakes.
We do have some real decisions to make. Being conscious that we could choose to live however we want, with no God to contradict us, is a great responsibility. What about psychopaths, for instance? Some people seem to have no appreciable conscience or in-built sense of compassion. Once we get good at detecting them, what should we do with them? Do we consider them "broken"? Can they be fixed? If not, is it moral to restrict someone's rights based on an innate compassion system failure, and the resulting high risk that they'll do harm to individuals or the society (especially if put into a position of power)? There may be no easy answers.
If it seems like I’ve got nothing but praise for scientists, it’s worth remembering the Manhattan Project. No less a scientist than Einstein promoted the idea of developing nuclear weapons technology. He later regretted it, but maintained the Germans would have developed them if the Allies had not. The pursuit of knowledge can always be thought of as an arms race. Once the knowledge is acquired, it’s time to make those hard choices about what to do with it.
Science and technology may soon open up new possibilities we've never had to consider before. Some world-changing technologies are nearly in our grasp. We are now able to genetically engineer radically different plants and animals, and could begin to change our own descendants with intention. (Some experts are calling for a ban on human germline mutation.) We could release a new species onto the planet and radically change the ecosystem almost overnight. We could replace Homo sapiens.
More modestly, and less controversially, we could cure a lot of diseases. Medical science has eliminated many causes of death, and is sure to eliminate many more. What looms largest on the horizon is the big challenge: aging. Can we cure old age? A machine can be kept going indefinitely if you maintain it properly, and replace failing parts when necessary. I, for one, want that longer life.
Dr. Aubrey de Grey, among others, believes we can solve the problem of aging and become effectively immortal. He, too, thinks of the entire body (brain included) as a machine made up of parts: cells, organelles within cells, DNA, and so on. He has identified seven categories of damage, effectively wear and tear, that the body inflicts upon itself as the inevitable byproduct of living. For example, there are "too few cells", "too many cells", "junk building up inside cells", and "uncontrolled replication" AKA cancer. Others have suggested seeking a permanent fix to the body’s design, so it stops damaging itself. This would be incredibly difficult because the body is so complex, and because we’d be fighting entropy. On the other end, much effort has been expended addressing diseases of aging once they manifest, but that’s impractical because aging manifests in hundreds of ways. Dr. de Grey suggests we should instead perform maintenance on the body: find ways to address those seven types of damage, reversing their effects. Do this upkeep periodically, and a person will remain, for all practical purposes, physically young and mentally fit.
If you question the approach’s practicality, that is fair. We can’t do this yet. But it’s hard to imagine it will never be possible, unless we drive ourselves extinct first, or -- again -- you believe there is something mystical about us. Some, though, object on moral grounds. Isn’t there something noble about aging, something important that is worth preserving? Isn’t aging necessary to society? If life were much longer, would it be robbed of its meaning?
Dr. de Grey has argued this line of thinking is a sort of collective learned helplessness. We’ve come to accept aging, even to celebrate it, because it happens to everyone. We accept it because we’ve never had a choice, before, to do anything about it. I would add, if you believe there is no life after death, then aging seems like a monster, ripping all our loved ones away from us. Of course we should try to develop the tools to thwart it, now that it might be possible. (Watch CGP Grey’s video The Fable of the Dragon-Tyrant to see this idea turned into allegory.)
Certainly, there could be downsides. What about overpopulation, will we keep having children at the same rate? Will we be able to find room and resources for everyone to live? What will it do to our ability to make progress on social issues, when old people with eighty-year-old viewpoints aren’t replaced by new generations? If a dictator assumes power, will their reign last for centuries?
We may not have answers to these, but a moral imperative remains to tackle the problem and turn aging into a choice our descendants get to make for themselves. If we choose not to develop the technology, we deny them that choice to use it.
While still struggling to understand ourselves and solve our human problems, our computers may be getting better at it than we are. We can analyze data, from Facebook for example, and know a person better than they know themselves. Once a computer is capable of making better life decisions for us than we are for ourselves, it will seem foolish, be foolish, to trust our own instincts.
Prof. Yuval Harari's Homo Deus contains some fascinating speculation about all this. For instance, we know the brain’s apparently unary sense-of-self has at least two components: one that experiences life in the moment, another that acts as narrator, stitching together a story of one’s life. It is the narrating self that we primarily think of as “ourselves”. The narrator judges a past experience based on its peak (the most intense moment) and its outcome; duration isn’t a significant factor. Given a choice between two unpleasant experiences, we will, in hindsight, prefer a longer one with a less-unpleasant end, rather than a shorter one, even though the shorter one holds less unpleasantness overall. This is deeply irrational when you think about it, and means we’re likely to make incorrect predictions about which choice we’ll be happy with. A computer could do a better job… even with the most important decisions, like what job to take, or who to marry.
If we are just physical machines ourselves, then the machines we build can, in theory, do anything we can. We've already managed to build computers that can beat the best players of Chess and Go. Computers can write music that is hard to distinguish from that of the greatest composers. They can drive cars. They can understand spoken language, and though their ability to parse the nuances of meaning needs a lot of work, you can count on it improving. This is, on one level, astounding. On another, we quickly adjust to the new toys, and take them for granted. The fantastical becomes commonplace overnight.
As we keep pushing back the boundaries of knowledge and technology, there's no reason to think -- absent a magical, spiritual element to humans that has yet to be repeatably demonstrated -- that our machines won't eventually become capable of doing anything humans can do. Shortly thereafter, our machines are likely to surpass us.
Do you doubt it? Nick Bostrom has an interesting observation here. We tend to view intelligence on a human scale, with “village idiot” on the low end, and a very smart human, like Einstein, on the high end. But our brains are only slightly better than those of our ape cousins. That tiny difference is the reason why we are running the world and they are in zoos. It may be hard to conceive of something we built being far smarter than us, but that, too, is a failure of imagination on our parts.
In Superintelligence, Bostrom explores how we might develop machines smarter than anyone alive. In fact, short of worldwide agreement not to do it, or a catastrophe that derails our scientific and technological advancement, it may be inevitable, merely a question of time. If you can accept that our brains are physical machines, then we can eventually solve the remaining mysteries of how our neurons work. Meanwhile our computers are increasing in capacity. Once we become able to scan a human brain in sufficient detail, and load that data into a sufficiently accurate simulation, the simulation would effectively be a human brain running on computer hardware. Once it's running inside a computer, it would be like any piece of software. It could be made to run faster... and copied. If you scanned an AI researcher in this way, you'd soon have a huge team of fast-thinking researchers who never got tired or made careless mistakes. They could cooperate to design a better system, and that system could quickly design a still better one, and so on. It might be mere weeks, or even less, until the newest system is smarter than the entire living human race.
This is only one way AI could be created. What follows is anyone’s guess, but there are some interesting theories.
Harari speculates that AI will give rise to a "useless class", a large proportion of the population with no economic value, because they have no skills that can't be done by a computer. Their labor won't be required once robots can do everything, including design and build better robots and repair damaged ones. We’re already seeing jobs lost to automation; we will see more. I posted about this recently so I won’t repeat myself here.
Perhaps we'll be wise, and arrange matters so a Universal Basic Income supports everyone. Or not. We could end up with a small group of elites controlling everything, while most of humanity fights over the scraps. That would be a terrible outcome, though not the worst possible.
Looking once more from the religious worldview, what happens in this world isn't what matters; it's all about the next. Logically, then, technology is at best a distraction or an amusement. It can't be important except where it intersects with religious views. On the other hand, if this world is all there is -- no Hell below us, and above us only sky -- then technology could prove staggeringly important, absolutely crucial. It could be the source of power to remake the world into a Heaven on Earth, or a Hell. There may be no God right now, but we could build one.
Nick Bostrom has made a Herculean effort to predict what might happen once we succeed at building an AI -- whether that’s ten, twenty, or eighty years from now. First, there is the distinct possibility of an intelligence explosion, as I describe above: the machine is smart enough to design and build its successor, possibly in days, possibly in minutes or seconds if it’s just a pure software change. Then comes the question of what the superintelligent AI will do when we give it an instruction.
A really brilliant AI, one that may surpass the cleverness and knowledge of all humanity combined, may still be stupid in ways we ascribe to the village idiot. It may not have self-awareness, self-preservation, a conscience, or any of our most basic values. Unlike us, it would not have evolved to believe in fairness. It would lack any value system we didn’t carefully program into it. It would be very, very good at solving problems, but otherwise, utterly alien.
Suppose such an AI belongs to a paper clip manufacturer, who simply instructs it to optimize for a high number of paper clips being made. The higher, the better. The machine may calculate that the best way to maximize its success is to take over all the world’s manufacturing resources, all its metal mines, and build robots that build robots that build paper clip manufacturing plants, miles across. In short order, our society could be wiped out, with all of us shouldered aside to starve beside mountains of paper clips. Having exhausted all of Earth’s metal, it would next build self-replicating spaceships that set out to do the same to every bit of the galaxy it could reach.
Wouldn’t we turn it off? Sure, if we saw all of that coming. But the clever AI would also understand our behavior better than we do. It’s like asking why apes didn’t stop us from putting them in zoos: they were outsmarted. The AI would predict we’d take steps to constrain it, and would design ways around that. At first, it would seem to be behaving purely beneficially, beefing up factory production in non-scary ways. The company’s board would be delighted with the balance sheets, and grant the AI increasing autonomy. While keeping up this appearance of safe compliance, the AI would simultaneously begin, surreptitiously, to put measures in place preventing it from ever being shut off. It might buy up more resources hidden under shell companies, then build backup copies of itself in secure bunkers, with power stations and automated defenses. More likely, it would execute some plan neither I, nor any human, could imagine. At some point, it would reach sufficient confidence that humans couldn’t prevent it from going full-bore. Then it would execute its true plan openly, and the mountains of paper clips would sprout everywhere. Nothing we tried at that point would stop it, since it would have predicted all our actions.
This is known as the “control problem”: how to ensure a thing that’s smarter than us will behave only in ways that benefit us. It is, to put it mildly, a bit of a poser.
Then again, if we're careful, it could be the last problem we ever have to solve. A properly designed computer-God, programmed to obey a value system that "maximizes human flourishing" (to borrow Pinker's expression), could correct all humanity's mistakes, end all human suffering, solve all the mysteries of physics. It could help us to spread across the galaxy, seeding a thousand worlds with humanity, and then billions. Perhaps every human could become an ageless god of a different planet, taking on projects that span millennia, creating new forms of life. Perhaps someday, whole new universes. Those people, our remote descendants, would never have to think: “This is not enough for me.” They would have us to thank for their supremely rich lives.
I’m deeply impressed by anybody who’s been patient enough to read all this. What started as loose musings on agnosticism has grown into a philosophical brain dump on rationality and an optimistic vision for the future of our species. I hope it will inspire someone else to let go of their particular piece of flotsam, and join the swim to shore.
Comments
Post a Comment