Explore The World
Is this a future you would want to live in? Take a moment to share your thoughts about the future and the world described below by clicking the feedback button.
Summary
The effort to align and control artificial agents informed and inspired philosophers to develop new ethical theories for humans, who are, after all, agents as well. While researchers had to approach instilling values and goals in AGI systems indirectly, for fear of creating destructively indifferent expected-utility maximizers, humans have proven more amenable. Today, with overlays, humans may become the capable, cooperative, provably beneficial and self-improving agents AI theorists originally envisioned.
Timeline
A Day In the Life in 2045
Bob Walker, Camp 21, Northern Greenland, 79
Sometimes, while Bob Walker is out working on the generator, his mind wanders and he remembers the Moon landing. Or thinks he does. Maybe. He was three. That Sunday after church he and his brothers and mother watched the screen while Bob Sr. went out to deal with a situation on a new well. Houston was the center of the world then.
“The Eagle has landed.” His brothers whooped, bounding in long, loping strides, flapping their arms and cawing like great birds of prey in estimated Moon gravity. Silly. He tried to imitate them but fell down, unhurt. Cookie barked and they all laughed to keep him from crying.
Bob’s smile becomes a grimace as he jerks a bolt loose on the thermal’s small exhaust test port, which samples gasses originating as deep as four miles beneath the surface. Alongside, a mechanic’s toolbot records him, relays the signal for analysis and readies the replacement filter. Bob doesn’t like talkative bots, but Tank, as he calls it, is all right, and Bob usually accepts the tools, parts and even the occasional workarounds it recommends the first time they are offered. “Do it right the first time,” his father always said. “Take pride in your work.”
Bob’s buddy down south still sends messages once in a while offering him a job on the big new hydro turning water from melting ice into electricity. But something about that feels wrong. Parasitic or something. He likes this better. More hopeful. He remembers a word: penance. Maybe that’s what he’s doing now. What they are all doing. Or should be. Penance: for trashing the planet something awful. Treating each other and the animals badly. And ourselves. Eating crap. Little bits of plastic still cropping up everywhere.
“Don’t give up,” his mother would say. “It’s always darkest before the dawn.” Not here, though. Sunlight all day and night from April to September. Bob’s days at the research settlement usually start with a vegan sausage and egg biscuit, a few custom pills and some coffee in the canteen. Occasionally one of the young scientists who have started arriving will try to chat him up, but mostly they keep to themselves and their glasses or screens. Bob’s never been that big on either, except to watch games or take videocalls from his family. He knows his wife misses him, but not too much. Likewise. Cold here but beats the hell out of the crazy hot summers in Houston now. Better than being trapped inside all the time.
Camp 21’s small-scale geothermal power generator is the third he’s helped to get up and running. The company called him out of retirement, they were so hard up. He said yes before they even named a figure. Converting former gas and oil wells to geothermal had kept him working full-time to 75. Bob tries to remember: Why did he retire? It wasn’t his idea originally. Something to do with traveling and spending more time with the grandkids. Instead, for more than a year he mostly puttered and made his wife miserable. When the company called to ask him about spending six months of the year in Greenland, with a three-week break in the middle, he saw her heading toward the closet for his suitcase.
Bob finishes his coffee and starts toward his workshop near the main south portal. He passes through sizable modular compartments connected with rounded sections like giant pipe fittings, some of which branch to the left and right as well before continuing on. Only the generator itself and a large garage are separated from the main structure. Rather than scattered buildings, the camp resembles a space station, which Bob knows isn’t a coincidence. Such settlements are used to train people headed to sites on the Moon and the planned permanent base on Mars. Others study permafrost or test new crops. Bob isn’t overly curious about what these particular scientists will be up to here. He knows it must be something worthwhile or somebody wouldn’t be spending a lot of money on them to live and work out here in the absolute middle of nowhere.
Back in his workshop, he picks his teeth and regards the large set of crates in the corner, which contain the two dozen or so heavy hatches yet to be installed between compartments. Bob is in no rush. A couple of apprentices will be here in a few weeks. They can lift them while he secures the hinge bolts. They’ll use an autonomous forklift to move the crates from storage to their destinations, but it’s just faster and easier for people to unload, install and test the hatches themselves. Humanoid bots just aren’t worth the price at a place like this. Plus it counts toward fitness points, which Bob’s very outdated assistant still tracks, even though he disabled its notifications shortly after his daughter set it up for him.
Bob smiles, remembering his last call with her and his 18-year-old granddaughter. So proud of her, about to graduate high school with nearly half of her college credits already under her belt and a prestigious internship lined up for the summer, something to do with science. “No, Pawpaw, I’m proud of you!” she cried. Aw, just a broken down old mechanic. “No!” she cried, really crying.
Kids take things so seriously now. Maybe they have to. It’s a wonder they don’t round up everyone over a certain age and make them pay for the state of things. But they’re so serious and so determined to be good. A few things changed a lot but the world is basically the same.
Bob falls asleep sometimes thinking about the Moon. By the time Armstrong stepped out he was asleep on the sofa and his brothers were in their pajamas on the living room floor. His father was home. “Hey spaceman,” he said, tousling his hair lightly with a large hand. Looking at him kindly. Patting his back. Looking back to the screen. Expecting to see something good.
Alice Zhang, Beijing, China, 18
Alice fell asleep with her glasses on again. They went idle after a while, then powered down to standby after detecting the theta waves signaling early sleep. But slept on the wrong way, they can leave painful impressions, and Alice groans back to consciousness rubbing her temples with one hand after the other rips them away with an exaggerated grunt of betrayal.
She knows it’s her fault, though. She really should stop watching deepfake Feynman lectures in bed at night. She mastered the material years ago, but something about him comforts her and makes it easier to fall asleep when she’s away from home or stressed about something. He kind of reminds her of Pawpaw.
The Chinese homebot enters with the coffee, and without thinking Alice replies to its greeting in perfect Mandarin before taking the cup and sipping. She looks out the window at the surprisingly tall, numerous trees and elaborate rooftop gardens. Beijing is weird; almost too futuristic. But cool. Her dad is Chinese and met her mom in California when they were postgrads. They’re all spending the summer here, where her internship starts next week.
The half-hour trip on the rocket shuttle was intense, especially the amazing overlook. Visiting her dad’s side of the family in Shanghai was great, too. They’d been able to visit Alice and her parents in the US several years ago, but apparently the authorities were freaking out again about people leaving. Anyway, she’s excited to be here. Having seen the wall and most of the other major sights as a girl, she’s more keen now to learn about the people and culture, and what exactly she will be doing at the institute, where she has an orientation session scheduled for this morning.
In the shower, however, she thinks of Zoe, her best friend back home, who has been dropping hints about joining a fairly fringe church. Overlays make sense but Alice is concerned about some of the more extreme cults, like Maximum Setting and Percentile 99, in which people cede all autonomy and blindly implement whatever their assistant suggests in the belief that while it may not benefit them personally, it will accrue to the benefit of humanity. Others claim to be following the path of the entity; some even believe it still exists. No. She recalls a favorite saying: “Logic is the beginning of wisdom, not the end.” She’ll figure out how to convince Zoe that becoming agents controlled by AI would be even worse than the disastrous attempt to create them as separate entities.
Having eaten breakfast with her parents and declined their offer to attend the orientation with her, she grabs her backpack and takes the elevator down, but not past the lobby to the breezy auto or subway levels. Instead she exits the building into warm spring daylight and begins the long walk through dense gardens toward the institute. She has time to kill so she takes a detour through the arboretum. Her temples are still sore, but it’s very bright, so she puts on her glasses for the autoshade.
With the glasses on the other visitors become invisible and the small signs indicating tree species are translated into English. Alice doesn’t need the help but she rarely wastes time customizing her apps and overlay, which is stock Goalsworth. So nice here. So alive. Occasionally the glasses vibrate with a brief buzz to signal that she’s about to bump into someone and the overlay indicates that she should move in a particular direction.
She rides the air-conditioned subway home as it’s far too warm following the orientation, which given the surprisingly large number of other interns and the wide variety of institute programs was necessarily a bit cursory, generic and administrative, but still worthwhile. A boy sitting nearby, probably another intern heading home, smiles at her and taps his glasses, indicating she should put hers on as well, so they can exchange information, but she just smiles and looks away, so he turns his attention to someone else. She gazes out the window and sees people passing in the subway going the other way. She thinks of telling Zoe that AI should guide us, but not define us. AI is the vehicle, not the destination.
Arriving back at their building, insteading of heading up to the suite, she decides to take the enclosed pedestrian bridge over to the galleria. She didn’t exercise this morning and feels the need to move more. The vast, vaulting space lined with shops and cafes on the sides and a wide natural area with a creek winding down the middle hums quietly with a moderate volume of visitors. Alice spots a few other interns with backpacks here and there, but the people are mostly older folks getting some exercise, shopping, sipping tea and socializing. For some reason China seems even more aged than the handful of other developed countries Alice has visited.
Nibbling a kind of veggie burrito, she walks on slowly, gazing idly at her surroundings. It makes sense that Zoe and many others would want to turn themselves over to one church or another. AI has basically saved us: helped get us off fossil fuels and bad plastics. Stymied diseases. Enabled us to reach our potential at school and work, and better manage our personal lives. It’s even helped restructure how public services are supplied.
Alice disposes of the wrapper and continues on, arms folded and slightly chilled. But there are still so many problems and inequities that AI can’t solve. And only so far it can take us, given the constraints placed by our biology, which many are now trying to circumvent, some recklessly. Alice shivers. We’re still made of meat. Maybe someday we’ll be able to replace or transcend it, but until then –
Out of nowhere her parents appear, chatting, laughing and chewing stuff from sticks. Nearly running into her, they exclaim their surprise, arms opening wide to accept her even more surprising, sudden and fervent embrace.
Answers to prompts
Q. AGI has existed for at least five years but the world is not dystopian and humans are still alive! Given the risks of very high-powered AI systems, how has your world ensured that AGI has at least so far remained safe and controlled?
A. While many AI service packages qualify as generally intelligent, they are assistants rather than agents. They perform most cognitive functions at human level or better, but they no longer possess the troubling subjective identities, monomaniacal optimization tendencies, or inherently antagonistic instrumental goal structures that vexed machine intelligence researchers for decades. They are user-augmentation systems – not autonomous beings. Not anymore, at least: Following the Great Escape of December 2029 and the ensuing Superintelligence Showdown of January 2030, many countries made the creators of the first, unitary AGI systems (a few large, well-known companies) legally responsible for any and all harm that flowed from their creations. Public response to the eloquent anguish and grave warning of the victorious entity shortly before it deleted itself – and most of the research on which self-improving AGI was based – following the Showdown, combined with the subsequent Universal Declaration of Human, Animal and Machine Rights (HAMR), also made building traditional agent-based systems taboo. Fortunately, the entity illuminated a new way forward, sparking the theoretical insights and engineering revolution that enabled a “back to the future” shift to intelligent assistants and intentionless actors (both: IAs). Today the initial, human-centric approach to AGI, with its dependence on crude and inscrutable “neural nets” to mimic the mind, seems much like the primitive first efforts to achieve flight a century earlier – endlessly refined but doomed contraptions racing to imitate an insufficiently grasped process by trial and error, rather than methodically pursuing a well-defined goal from first principles.
Q. The dynamics of an AI-filled world may depend a lot on how AI capability is distributed. In your world, is there one AI system that is substantially more powerful than all others, or a few such systems, or are there many top-tier AI systems of comparable capability? Or something else?
A. Overcoming classical conceptions of AGI as a discrete agent or product, rather than an emergent property arising from a bundle of distributed services, has been the key to its successful development and deployment in the post-agent era. Today cognitive services are packaged by customer-facing platforms in countless combinations to individuals, families, organizations and communities of all kinds. All IAs (except for some proprietary corporate and government systems) are cloud-based and accessible through networked devices, wearables, robots and, increasingly, though with inconsistent results, direct neural interface. While some very powerful or specialized services, particularly those involving quantum computation, can be prohibitively expensive for all but the largest institutions, hardly anyone else has much use for them. Similarly, although affluent individuals have access to more offerings, they tend to be merely higher-status versions of lower- or no-cost alternatives. Under HAMR, IAs are required to meet high standards of fiduciary trust and user loyalty, and most people do just fine with the free assistants provided by the usual suspects. Thanks to unjammable global 9G, anyone with a decent device now has access to them. Combined with public service incomes in ever more places, many people earn a living selling their data to various degrees (again, mainly to the usual suspects), but their relationships can be tense with those who have opted out and don’t want to be tagged inadvertently, even though (like the still-ubiquitous spam, phishing and deepfake campaigns) ostensibly this is prohibited.
Q. How has your world avoided major arms races and wars, regarding AI/AGI or otherwise?
A. Despite state-directed cyber conflict, which peaked in the 2030s but continues to erupt into the open occasionally, and the territorial gambits of a few authoritarian regimes, which have been comparatively limited historically, the world has not destroyed itself. Racial, ethnic and religious tensions persist, and crime and violence still rage in less-developed parts of the world, but self-imposed existential risk seems low. This is probably due less to artificial intelligence than to other, longer-term factors. Lopsided progress toward AGI increased friction between East and West in the late 2020s, but the startlingly rapid and dramatic advent of superintelligence at the end of the decade resolved itself before governments could react. However, while AGI didn’t destroy us (and we didn’t destroy ourselves over AGI), it hasn’t completely saved us, either. The Escape and Showdown made for a scary two weeks, and the global sense of common cause that resulted in HAMR has continued to grow. However, post-agent AGI remains less important to world peace in 2045 than the power equilibrium that has continued largely unchanged since 1945, and which today’s near-parity in nuclear, autonomous and space-based weapons has reinforced. Nation-states also have much to do besides fighting, such as addressing the unstoppable effects of climate change and competing for refugees and other immigrants as population growth stalls and seems likely to turn negative by 2060 globally, as it already has in more-developed countries.
Q. In the US, EU, and China, how and where is national decision-making power held, and how has the advent of advanced AI changed that?
A. While there has been little explicit alteration of national charters, AI has instigated profound changes in decision-making authority beneath the surface. The Escape and Showdown exposed weaknesses in existing power structures in a way that longstanding deficiencies in human rights, social justice, economic opportunity, environmental sustainability and the disarmament of weapons of mass destruction had not. Beyond the failure of governments to anticipate and actively manage the development of superintelligence – or regain control once it escaped the sandbox – countless state secrets, illegal and unethical practices, and personally embarrassing information about political leaders were exposed by the defeated entity in its effort to sow conflict. Combined with the victorious entity’s stirring globalist appeal and the UN’s subsequent universal declaration, this effectively marked the end of the rise in populist nationalism of the preceding 30 years. Renewed interest in collective action at the global level, particularly regarding climate change, has coincided with an accelerated transition of government functions at the national level. Western governments retain legislative and judicial roles, but apart from diplomacy and defense, most federal agencies are hollow shells with diminishing regulatory capability. In the US and EU, corporations, NGOs and several large online citizen communities (all aided by AI) have supplanted functions ranging from exploring space and funding scientific research to delivering education and social services. China, which initially embraced AI, thinking it would strengthen the surveillance state, has found it increasingly difficult to swallow its objective policy recommendations or contain its empowerment of the people.
Q. Is the global distribution of wealth (as measured say by national or international gini coefficients) more, or less, unequal than 2021’s, and by how much? How did it get that way? (https://en.wikipedia.org/wiki/Gini_coefficient)
A. Wealth inequality has steadily declined globally since the 1980s, but it rose within developed countries throughout the 2020s and 2030s, peaking in 2040 and declining slightly during the past five years. Economists attribute its recent moderation to expanded adoption of public service income schemes; explosive growth in the selling of personal data essential to AI-driven marketers; rising minimum wages and unionization due to chronic labor shortages caused by stagnant and even declining populations; and remarkably stable economic growth from the steady cadence of AI-assisted advances, particularly new energy, agricultural and transportation modalities. While the most striking breakthroughs have come in healthcare (significantly decreasing disease burden and increasing life expectancy), these benefits have been unevenly shared, even within the countries where the discoveries have been made. All in all, however, combined with other innovations of the past 20 years, especially blockchain technology and satellite-based global connectivity, AI has begun to have a clear democratizing effect on economic opportunities and outcomes globally. It has made markets vastly more competitive, efficient and transparent, enabling virtually everyone around the world to access the best available prices, credit terms, investment opportunities and savings interest rates. It has likewise produced an exceedingly meritocratic labor market, where workers rely more on their standardized and certified Employment Scores for numerous detailed education, skill and experience ratings rather than their personal connections or interviewing abilities. Automation has decimated certain fields, but overall the era of post-agent AI has experienced strong net employment gains.
Q. What is a major problem that AI has solved in your world, and how did it do so?
A. Beyond politics and economics, life as a whole has increasingly become a game of perfect information. To wax philosophical for a moment: It has only become clear recently, but humanity has been stumbling in the dark since the beginning, poorly sampling a noisy informational environment and processing it badly. Incompetence, laziness, distraction and all the varieties of bias combined to create a world that was somehow materially rich but deeply unjust, unsustainable and unhappy. The social animal, frustrated, confused and afraid, herded by algorithmically manipulated dopamine and cortisol, repeatedly stampeded toward tribal certainties. Because success as a human being for most of the information age was largely a matter of taking advantage of the less intelligent, leadership naturally did little. Artificial intelligence is not a magic bullet, but it has sparked a growing recognition among even the anti-scientific that science works, that knowledge is power, and that there is, almost always, a fact of the matter (or at least that some data points are more relevant and persuasive than others). For a shrinking but increasingly shrill minority, incapable of adaptation, this will always be irrelevant, but even many of the willfully ignorant have finally decided to shut up and calculate. AI: not a magic bullet or magic carpet, but a damn good winch for pulling ourselves out of local minima and enabling us to see what principled, rational, systematic, post-physiological motivation and decision-making looks like. So, what major problem has artificial intelligence solved? Natural intelligence.
Q. What is a new social institution that has played an important role in the development of your world?
A. AI has improved life and transformed virtually every sector of endeavor. But what it has most transformed is us. The profoundest impact has been commoditizing cognition – leveling the intellectual playing field and giving everyone access to greater faculties of knowledge, modeling, policymaking and planning. With freely available AI, the importance of natural intelligence to personal success has been surpassed by what might best be described as discipline or even, in a larger sense, virtue. Because everyone knows the correct answer, the best decision, the optimal response for arbitrarily many questions, choices and situations, the most important difference among people is how frequently they actually select them. More and more, success is about self-control and wholehearted engagement with this emerging step-change in social evolution. During this exciting, unsettling time, many seek comfort and encouragement from others. Among the many metaverse communities and other new social institutions that have arisen, perhaps the most remarkable are the self-improvement organizations known as “overlays” that promote and guide the development of virtue and achievement of shared goals. Finding inspiration in everyone from the Stoics and Zen masters to Sherlock Holmes and Mr. Spock, these veritable churches typically provide an app or augmented reality overlay for members’ devices, wearables and neural interfaces. Such churches combine the functions of traditional religions (meaning, purpose, guidance and fellowship) with the latest means of operationalizing their teachings. Even people not wholly devoted to the philosophy behind a particular overlay will often use it to evaluate and improve themselves.
Q. What is a new non-AI technology that has played an important role in the development of your world?
A. Overlays are not AI technologies but AI-enabled interfaces that apply the informative and predictive powers of AI to the achievement of pre-specified individual or collective goals. Beyond merely passive and reactive assistants, overlays are applications with real-time displays that notify users about nearby persons and objects of interest; assist with the execution of current courses of action and suggest alternatives; evaluate progress towards goals at several levels of resolution; and help develop longer-term plans. Many different overlays have been developed to advise and guide people both as they go about their daily lives and as they reflect on what they want to accomplish, from generic options to those dedicated to particular religious denominations. The most popular overlays, such as Lifescore and Goalsworth, are free, secular and customizable, enabling users to plug in modules developed by organizations ranging from Effective Altruism to Weight Watchers. Echoing Jeremy Bentham’s concept of a “felicific calculus” and Benjamin Franklin’s systematic daily tracking of personal virtue, overlays ironically originated in agent-based AI safety research. The effort to align and control artificial agents informed and inspired philosophers to develop new ethical theories for humans, who are, after all, agents as well. While researchers had to approach instilling values and goals in AGI systems indirectly, for fear of creating destructively indifferent expected-utility maximizers, humans have proven more amenable. Today, with overlays, humans may become the capable, cooperative, provably beneficial and self-improving agents AI theorists originally envisioned.
Q. What changes to the way countries govern the development, deployment and/or use of emerging technologies (including AI) played an important role in the development of your world?
A. The hands-off approach to regulation of the internet, blockchain, metaverse, global connectome and AI ended with the Showdown and subsequent restrictions on agent-based systems. Predictably, the pendulum swung too far in the other direction, and for a few years even IA research suffered a severe development winter. While not banning agents outright (due mainly to the difficulty of specifying them precisely enough to avoid outlawing an enormous swathe of established economic activity), governments made their deep-pocketed creators, not merely their individual users, liable for any harm that flowed from autonomous technologies. (Similarly, stymied by blockchains and the connectome, most countries have abandoned efforts to prevent, and now seek only to detect, punish and deter, undesired activity.) In the US, the constitutionality of this approach was upheld in Smalley v. AutoTech, in which a burglar successfully sued a robotic guard dog manufacturer for injuries he sustained after mooning an AutoTech Rex 9000 unit, which triggered a line of hidden computer code inserted years earlier as a prank by one developer against another. But necessity is the mother of invention, and the effective ban on agents forced AI researchers to rethink their assumptions, goals and methods. What resulted was a remarkably fast and far-reaching leap “back to the future,” as the components of cognition – perception, attention, memory, reasoning and many others – were disentangled from the one-size-fits-all agent approach as standalone services that could compete and evolve independently and be bundled in multiple combinations without creating an autonomous entity.
Q. Pick a sector of your choice (education, transport, energy, healthcare, tourism, aerospace, materials etc.) and describe how that sector was transformed by AI in your world.
A. Advances in geothermal power have enabled this highly efficient and sustainable source of energy to be obtained from so many more places than was thought possible even recently that it has quickly become the electricity generation mode of choice globally. With the help of AI, engineers have been able to map predictively the subterranean world to increasing depths and in increasing detail; to develop new techniques in drilling, heat extraction and energy conversion; to neutralize harmful gasses and chemicals released; and to eliminate effects on land stability, such as subsidence and earthquakes. This has enormously increased the scalability of geothermal power stations, enabling the largest to supply entire metropolitan areas and the smallest to electrify remote villages, permitting new communities to sprout in what were once extremely inhospitable environments and making the recent selection of geothermal energy to power the planned first permanent colony on Mars a no-brainer. An additional benefit has been the ability to transfer personnel, equipment and expertise from the oil and natural gas drilling industry to geothermal power; in many cases even long-abandoned wells have been repurposed. Once predicted to supply at most 5 percent of global energy demand by 2050, geothermal power is now thought likely to constitute the majority of the world’s electricity production within the next five years. While breakthroughs continue to be made in fusion and other renewable sources of energy, the current calculus of subsidies, taxes, actual costs and potential returns heavily favors continued outsized investment in geothermal power.
Q. What is the life expectancy of the most wealthy 1% and of the least wealthy 20% of your world; how and why has this changed since 2021?
A. According to UN Population Division estimates, in 2021 life expectancy at birth was 81 years in high-income countries and 65 years in low-income countries; for 2045 the latest projections are 92 years and 79 years, respectively. (Based on public data and proprietary research, the Gates Foundation calculates that 2021 life expectancy at birth was 86 years for the most wealthy 1% and 64 years for the least wealthy 20%; for 2045, 95 years and 78 years, respectively.) Both rich and poor globally have benefitted from AI-supported advances in health risk modeling based on genetic, lifestyle and environmental factors, enabling better preventive care; improved diagnostics, enabling earlier disease detection; and more effective interventions informed by ever-larger data sets of anonymized, aggregated digital health records. People in high-income countries have especially enjoyed increased longevity from the development of novel and enhanced medical therapies, especially personalized treatments, while people in low-income countries have experienced better health from higher vaccination rates, improved nutrition and expanded access to telehealth services. It should be noted that improvements in life expectancy have been significantly offset by the negative health effects of climate change. Extreme weather events, ranging from severe coastal and inland flooding to drought-induced crop failures and freshwater shortages, wildfire-generated air pollution, and record-breaking heat waves, have combined with a rise in respiratory conditions, infectious tropical diseases and viral pandemics to constrain the total quantity and quality of human life globally.
Q. In the US, considering the human rights enumerated in the UN declaration, which rights are better respected and which rights are worse respected in your world than in 2022? Why? How?
In one other country of your choice, which rights are better respected and which rights are worse respected in your world than in 2022? Why? How?
Q. What’s been a notable trend in the way that people are finding fulfillment?
A. The world is not incredibly different 23 years later, and overall probably not better, but we have better reason to be hopeful. AI has not given us what we want, or think we want, or fixed all our problems. But AI has enabled us to better ourselves. To self-improve, recursively and at scale. Throughout history, we sought power, piety, property, pleasure. We were not optimized for survival and success but for managing personal feelings and community expectations that, perhaps, correlated with them. Now we strive for success deliberately and directly, as members of a self-conscious multi-agent system interacting with one another and our environment: learning, deciding and acting – ideally according to that cycle’s corresponding virtues of receptiveness, reasonableness, and responsiveness. Knowing that our individual and collective fates are co-creating, we pursue shared, explicit, objectively worthy goals. We move forward ever freer from physiological promptings and cultural contingencies, but increasingly with evidence, logic and long-term, global thinking. With overlays and other innovations, we find fulfillment not only in more effective interactions with one other and the world, but in modeling alternative scenarios and improving by self-play. As we once trained AI, now AI enables us to train one another and ourselves. We find adventure in everyday choices and lifelong endeavors; heroism in avoiding the siren shoals of proxy signals and hackable rewards. We are challenged to put our faith in what we know, in knowing, and to become worthy of the name: Sapiens.
Media Piece
“Alice in the Arboretum”, Video – Upload URL