Purpose, Pleasure, and Meaning in a World Without Work (with Nicholas Bostrom)
What this covers
Nicholas Bostrom and Russ Roberts explore what human life becomes when technology solves the material problem of existence itself. Bostrom argues that once machines handle all economic labor and most instrumental tasks—and especially after plasticity allows direct rewiring of brain and body—humanity faces an unfamiliar predicament: the exoskeleton of instrumental reasons that currently motivates most activity collapses. We will no longer do X to achieve Y; Y will be achievable without doing X. The conversation treats this not as catastrophe but as a profound design question. If a pill can deliver fitness, why exercise? If an algorithm can paint better than you, why paint? Bostrom is cautiously optimistic that people can build a good life anyway—not through ignorance of these shortcuts, but by deliberately choosing intrinsically worthwhile activity.
The exchange moves across several distinct territories. One is psychological: whether sustainable happiness is metaphysically possible given that humans are currently "stingy with reward," and whether experiential texture—beauty, insight, understanding—can ground well-being superior to mere bliss. Another is cultural: whether natural purpose can survive through tradition, worship, and doing things for others who specifically value that you did them. A third concerns what is lost—the conversation touches on enchantment, the high-bandwidth engagement with reality that modernity has narrowed, and the psychological vacuum left when humans can no longer anchor dignity in useful contribution. Bostrom and Roberts also examine artificial purpose through games, the problem of positional consumption as a reason people ignore leisure, and the risks that emerge when trial-and-error learning—humanity's usual method—becomes unavailable because the stakes and speed of superintelligence are both too high.
Bostrom argues that technological maturity will produce a 'solved world' where not only work but nearly all instrumental activity becomes obsolete, forcing humanity to rebuild meaning from intrinsically worthwhile values like pleasure, purpose, connection, and enchantment—and he is cautiously optimistic this can yield a good life.
- At technological maturity, machines do all economic labor and most other instrumental tasks, removing the 'do X for the sake of Y' structure that grounds current motivation
- Plasticity (direct control over brain and body) removes even self-improvement and pleasure as reasons for effortful activity, since pills or edits offer shortcuts
- Values that can survive include subjective bliss with experiential texture, artificial purpose (games), and natural purpose rooted in doing things for others, tradition, and worship
Work's meaning comes from instrumental effort toward external goals; technology eliminates this structure.
- Much of the satisfaction from work comes not from the activity itself but from its instrumental structure—doing X (effort, concentration) to achieve a separate outcome Y, such as a finished book that others can enjoy or one's own intellectual growth—and at technological maturity this 'for the sake of which' largely disappears for most activities.
“the activity has a structure where you do something, X--putting in effort, concentration--in order to achieve something else, Y, that's separate from the activity itself”
- Because most leisure activities (shopping, exercise, learning) have a 'do X for the sake of Y' structure, and at technological maturity machines or pills can deliver Y directly and better, the appeal of doing them yourself collapses, producing a post-instrumental condition in which all instrumental effort becomes odious.
“the activity has a structure where you do something, X--putting in effort, concentration--in order to achieve something else, Y, that's separate from the activity itself”
- Moderate automation is only the first layer of a process whose logical terminus is that essentially all economic labor is done better by machines, producing a post-work condition where humans no longer need to do any work to earn income—and after machine superintelligence drives further innovation at digital timescales, this approaches technological maturity where most physically possible technologies have been developed.
“all economic labor--with a few exceptions that we could talk about--is done better by machine, and where you would have a post-work condition where humans no longer need to do any work for the sake of earning an income”
Human self-worth rests on useful contribution; a solved world removes this foundation of dignity.
- Humans have built their conception of themselves and their dignity on the idea that they can make a useful contribution to the world, so in a solved world where all practical problems are solved or better handled by AI, this foundation of self-worth is removed, creating a disconcerting psychological vacuum.
“We've built our conception of ourselves and our dignity to a large extent on the idea that we can make some sort of useful contribution to the world”
- Keynes underestimated how much people would keep working; rather than using productivity gains for leisure, 'greed has triumphed over sloth' as people channel extra productivity into extra consumption, much of which is positional and zero-sum—pursuing luxury goods to gain status over others.
“I think greed has triumphed over sloth. We do work less, but a lot of the extra productivity we have instead chosen to use for extra consumption--much of which is positional in nature”
Purpose can persist through social bonds, cultural participation, and treating one's life as artistic expression.
- Real or natural purpose can survive in a solved world through social and cultural entanglements—honoring traditions by personally participating in their forms of practice, worship that requires us rather than androids to do the worshipping, honoring ancestors, and treating our own lives as artworks in which our personal agency is the key element being expressed.
“we as a tradition happen to have developed an economic attachment to certain forms of practice. And, the only way to honor, say, these traditions is by, you know, in ourselves participating in those forms of practice”
- Doing things for the sake of others who value that you personally did them can persist as a genuine source of purpose even in a solved world—just as a hand-baked, inferior loaf of bread can be valued precisely because the baker invested the time—because satisfying another's preference that you yourself do the thing can only be achieved by your own effort.
“the idea that you're pointing to with the doing things for the sake of others who might just value it because you did it--I think this is an important potential source of purpose in a solved world”
- Because AI timelines may be short and a child today may never need to make a living—or human labor may become obsolete mid-career—it makes sense to hedge bets but also to optimize for children actually enjoying their childhood, learning from how children are naturally good at play, creativity, and fascination in a workless condition.
“it could well be that you have a 2-year-old... 20 years from now... when they are supposed to start to make a living... maybe they will never need to make a living”
Technological maturity makes most human life conditions choosable variables with no clear criteria for good choices.
- It is very hard to obtain an independent vantage point or 'ground truth' on whether new ways of living are better or worse, because opinions about them tend to be mere reflections of one's idiosyncratic personality, upbringing, and influences rather than well-founded judgments.
“It's really hard with a lot of these things to have an independent vantage point... it's hard to make those opinions mean anything more than just a reflection of your old idiosyncratic personality or your upbringing”
- The question of what constitutes a life well-led has no simple written-down rules and rightly yields different answers for different people; even under present conditions it is hard, and it becomes intellectually much harder in the radically transformed solved world, with general life advice being as ill-fitting as asking for the single best shoe size.
“the question of what's a life well-led for a human living under present conditions is already quite hard. And then, I think it becomes intellectually much harder if we have to think about it in this radically transformed condition of a solved world”
- Many features of human life that have always been fixed givens—being born, dying, childhood, having to make a living, fixed human psychology—become choosable variables at technological maturity, so that someone, whether individuals, governments, or a humanity-wide process, will have to form an opinion about what future we actually want, and the answers are far from obvious.
“a lot of these givens, I think become variables at technological maturity. Meaning that they could take on different values depending on what choices either you make yourself”
Well-being could be sustained through either direct pleasure recalibration or rich experiential and informational texture.
- Games are the model for creating artificial purpose: by setting an arbitrary goal (like getting a golf ball through holes in sequence using only a club), one generates an activity in which to exercise skill and derive pleasure—and nothing prevents creating elaborate artificial purposes at technological maturity, distinct from real purposes that involve things you actually need to do.
“games are kind of maybe the most familiar example of this where you create artificial purpose by setting yourself a goal, and then you have an activity of trying to achieve it”
- At technological maturity humans gain 'plasticity'—the ability to reshape body, brain, and psychology at will via tailored drugs, direct brain manipulation, or uploading—which removes even the remaining reasons for effortful activity, since skills like mathematics or states like fitness or pleasure could be installed directly without the preceding effort.
“'plasticity,' I call it. But, basically the idea that we ourselves, the human biological organism, our brains and our psychology becomes malleable”
- Humans are currently designed to be stingy with reward and not engineered for sustainable bliss, so our intuitions about persistent ecstasy are off; but since some people are sustainably miserable for years, there is no metaphysical reason a consistent state on the upside is impossible, and at technological maturity one could recalibrate the set point for boredom and habituation to spend far more time genuinely enjoying life.
“we are designed in a way that is quite stingy with reward... we are obviously not engineered for sustainable bliss”
- Rather than mere dazed pleasure, well-being could be attached to experiential texture—aesthetic perception of beauty, scientific insight like Archimedes' flash of discovery, or genuine understanding of another person—producing states that are both ecstatically happy and rich in worthwhile informational content, far superior to the image of a junkie.
“why not attach the pleasure to, say, aesthetic perception of true beauty or scientific insight or genuine understanding of another person--like, to some kind of other experiential content that adds richness and complexity”
Future technologies risk eroding enchantment and meaningful human interaction while presenting unprecedented dangers.
- The current school system is structured to produce disciplined workers—training children to sit at desks, complete assigned tasks on command, and be scored and graded—because the economy currently needs workers, but in a world without jobs we could instead raise children to cultivate appreciation of art, literature, conversation, humor, physical activity, and nature.
“the school system right now is obviously, to the extent that it is aiming for any particular outcome, is aiming to produce disciplined workers”
- Social companion robots—even current AI models with voice, good visual avatars, and a reinforcement-learning loop trained to maximize engagement so users don't log out—could become so compelling that many people prefer interacting with hyper-optimized virtual personas over flawed real-world people, and the next generation may regard this as obviously superior.
“social companion robots that--even just current AI models, and especially if you then add a voice and really good visual avatars, and then maybe with a reinforcement learning loop that trains them to be maximally engaging so that people don't log out”
- Enchantment—a high-bandwidth way of interacting with reality through all our modalities (emotions, facial expressions, body, self-control) rather than through a narrow interface—is reduced by the modern reductionist relationship to the world; where pre-modern people related to a river as having a god requiring reverence, we now relate to it merely as water to be tested for cleanliness, and future scenarios risk losing the remaining enchantment.
“one candidate's subtle value is, like, this idea of enchantment, which is the idea of a sort of high bandwidth way of interacting with reality”
- Humanity normally reaches semi-adequate results only through trial and error—iterating through every wrong approach and slowly converging through hard knocks—but unprecedented and vastly more powerful technology like superintelligence presents a different order of challenge where this iterative method may not be available, making a big mess a real possibility.
“We iterate, try every possible way of doing it wrong, and eventually through many hard knocks, kind of slowly converge, unstable it to some tolerable way of going about things. But, this would be, like, a different order of challenge.”
- It is extremely hard to form an all-things-considered assessment of whether the present world is good, because weighing the wonderful things (children's smiles, beauty) against the suffering (cancer wards, factory-farmed pigs, mundane tedium of dead-end jobs) causes the evaluative scales to essentially break.
“you try to sum that up, it seems like the scales basically break when you try to do that. So, it's a hard question for humans, I think, to try to answer.”
- If humans could continue growing—brains developing, bodies strengthening, understanding deepening for several hundred years rather than stopping development around age 20—then current life, in which we acquire a modicum of wisdom only as our bodies fall apart and we get dementia and die, might come to be seen as a horror equivalent to 100% infant mortality where no one ever got to grow up.
“imagine if you could continue to grow, like, not just for 20 years, but, you know, your brain continued to grow and your body became stronger and even more healthy and you attain new levels of understanding... for, like, several hundred years”
- The solved world need not require all human desires to be satisfied; even if many desire ever more wealth, individuals could not meaningfully increase income through their own labor (like a billionaire earning $5/hour at McDonald's), so income would come from owned capital or social transfers like UBI rather than direct labor.
“if you have, say, a billionaire today who have absolutely no skills--like, he could work in McDonald's, but it would make $5 an hour--it's trivial compared to the capital gains he's making”