97 | John Danaher on Our Coming Automated Utopia
What this covers
John Danaher, a legal scholar and philosopher at the National University of Ireland, sits down with Sean Carroll to argue that automation will soon render human labor obsolete—and that this is fundamentally a good thing. The episode traces three interconnected claims: that robots and AI are steadily displacing workers with no offsetting job creation (drawing on empirical work by Acemoglu and Restrepo showing net job losses in manufacturing); that the conditions of work have deteriorated for most people through fissuring of corporations, precarious work, and income inequality; and that humans need not suffer this loss because work is not the only source of meaning. Danaher proposes that in a world where material needs are met through basic income, people can derive the goods work currently supplies—mastery, community, contribution, recognition—through freely chosen activities, which he frames using Bernard Suits's philosophical definition of games.
The conversation moves across several distinct territories. It examines how Moravec's Paradox and the polarization effect reshape labor markets, how automation has historically operated through both substitution and complementarity effects, and whether those historical patterns will hold as AI advances into cognitive work itself. Danaher distinguishes between jobs (bundles of tasks organized for economic reward) and work (the condition of performing them), arguing technology automates the former while leaving the latter's bad conditions intact. He addresses several objections head-on: that remaining jobs offer reassurance (they mostly don't, falling into superstar markets or undesirable care work), that games feel hollow because their rules are arbitrary (but football's rule evolution shows we discover meaning through play), and that humans will grow bored (chess has sustained interest for millennia despite machine mastery). The episode also sketches an alternative cyborg utopia where humans become more machine-like to compete, leaving the virtual utopia as the more livable path forward.
Danaher argues that automation is steadily rendering human labor obsolete, that the loss of work is desirable because work's conditions are increasingly bad, and that humans can find meaning in a 'virtual utopia' of freely chosen game-like activities rather than by competing with machines.
- Technology automates the tasks that make up jobs and empirical data suggests robots produce net job losses rather than neutral or positive employment effects
- Work increasingly undermines freedom, is precarious and unequal, and most people report dissatisfaction with it
- Meaning currently derived from work (mastery, community, contribution, recognition) can be obtained through chosen games once material needs are met
Automation displaces workers faster than new complementary tasks emerge to reabsorb them.
- Work is not any particular activity but a condition under which humans perform tasks in return for economic reward, and a job is a socially or economically defined role made up of a bundle of tasks; this distinction matters because technology automates tasks rather than whole jobs.
“work isn't any particular activity or task; it is rather a condition under which humans perform tasks. So it's a little bit of an abstract notion. And a job then is effectively a socially or economically defined role that is made up of a bunch of tasks”
- Per Moravec's Paradox, high-level cognitive tasks are often easiest to automate because they can be reduced to algorithms, whereas physical manual labor is hardest because it relies on unpredictable stochastic variables; the modern reframing is that routine work is easily automatable while non-routine work is harder, though new AI is increasingly handling even non-routine tasks.
“high-level cognitive jobs are oftentimes the easiest things to automate, because they tend to be very easy to formulate an algorithm to perform those tasks. Whereas physical, manual labor is often the most difficult thing to automate”
- Historically automation has disrupted jobs without causing structural mass unemployment because, alongside the substitution effect where machines replace human labor, a complementarity effect opens up new tasks in which humans have a comparative advantage, allowing workers to redefine their roles around different tasks.
“there are complementary tasks humans can perform alongside machines that kinda opens up a whole vista or new space of work for the future. And this is why we haven't seen this trend towards kind of structural unemployment over the long term as a result of technology”
- Empirical work by Acemoglu and Restrepo on US labor markets from 1990-2007 found that each additional robot displaced between three and six workers with no offsetting net gains across other industries, a similar effect was found in French firms, and this occurred during a period of relatively low robot uptake, suggesting larger losses as robotization increases.
“for every robot a company uses in the US it tends to displace either between three and six workers. So it was a kind of a net loss of employment. It was data based on from 1990 to 2007”
- Since the 1800s, agricultural employment in most European economies and the US has fallen precipitously from the majority of workers to less than 5% of humans, illustrating the long historical trend of task automation reducing labor in a given sector.
“the majority of workers worked in agriculture in the 1800s up until the early 1900s, and now we're talking about less than 5% of humans employed in agriculture”
- Per David Autor's work, technology produces a polarization effect on the labor market by automating middle-skill middle-income jobs, pushing people into either low-skill low-income work (often arduous physical labor with health repercussions) or high-skill high-income work that has higher barriers to entry, creating inequality in both income and the meaningfulness of work.
“he talks about this polarization effect on the labor market as a result of technology, that people are pushed into either low skill, low income work, or high skill, high income work”
Work conditions are deteriorating for most workers through fissuring, outsourcing, precarity, and inequality.
- Danaher's critique is not that work is intrinsically bad but that the conditions under which we perform jobs are getting worse for the majority of workers, through fissuring of the workplace, more outsourcing and precarious gig work, growing income inequality since the 1980s, and polarization into low-skill and high-skill brackets.
“the conditions under which we perform work, perform our jobs tend to be getting worse for the majority of workers. That's not to say that there aren't people who benefit from the current system, there clearly are. But the majority seem to be losing out”
- Since the 1970s, large corporations that once employed diverse workers under one umbrella have shifted to focusing on a core competency and outsourcing the rest to contractors, because this fissuring increases returns to shareholders and reduces costs to consumers while workers get the worst deal, and surveillance technology has enabled it by allowing remote enforcement of corporate standards.
“companies are focusing on what they view as their core competency, and then they're outsourcing a lot of the other forms of work, to part-time contractors or other companies”
- Even though average hours worked per year may have declined, an increasingly competitive and precarious labor market means work is colonizing our lives, with people spending far more time and attention up-skilling, making themselves employable, and worrying about employment, illustrated by a 12-year-old who took a class mainly because it would look good on his CV.
“even if we're not spending more time at work, and there's some evidence to suggest that the amount of hours that people spend per year working has gone down, we're spending a lot more time thinking about work and caring about work”
- The degree of attachment and nostalgia a person feels for work is a function of the kind of work they do; high-skill workers rewarded for creative problem-solving and granted autonomy are most attached to the current system, whereas many people already derive most of their satisfaction and meaning from non-work activities like family, community, and hobbies.
“the level of attachment and nostalgia you have for work is probably a function of the kind of work that you currently do”
- The physical trading floor of stereotype, with people barking buy and sell orders, has largely disappeared; most trading now happens digitally, with more than 50% of all trades executed automatically by algorithms, and the New York Stock Exchange floor has been shrinking since its 1990s peak.
“most trading activity is performed by algorithms. The estimates are a little bit tricky here, but most people will assume on pretty good grounds, it's more than 50% of all trades are executed by algorithms automatically”
- Gallup's recurring global workforce surveys over roughly 15 years consistently find that most people express dissatisfaction with their work and do not feel fully engaged by it, partly driven by the competitive and precarious nature of the modern workplace.
“one of the consistent findings from their surveys over the past 15 years or so, is that most people express dissatisfaction with the work that they're doing. They don't feel fully engaged by it”
A low-work future where people play meaningful games can supply mastery, community, contribution, and recognition.
- Even if the overall landscape of games is finite, individual games contain near-limitless possibility spaces—people have played chess for thousands of years and still find it fascinating even though machines beat humans—so a low-work future is unlikely to leave humanity bored within any meaningful human time horizon, though at the infinite limit boredom would set in.
“People have been playing chess for thousands of years, and a lot of people still find it intrinsically fascinating, that there's lots of room left to explore in it, even though machines are better than humans at it”
- Counterintuitively, humans have always been a virtual species, using technology to build insulated environments shielded from real-world effects that increasingly conform to our desires; the apotheosis of this trend is realizing that much of life consists of socially constructed, game-like rules over which we can gain control.
“humans have always been a virtual species... we've always used technology to create insulated environments or niches in which we can survive that are hidden or shielded from a lot of real world effects”
- Using Bernard Suits' definition of a game as a goal-directed activity guided by arbitrary or contingent rules, a future in which work is optional could let people freely choose which games to play, and these games can supply the non-income goods of mastery, social contribution, community, and recognition, offering a recipe for mass meaning rather than meaning reserved for an elite few.
“the future I'm imagining, this virtual future, is one in which you actually do have a choice over the kinds of game that you get to play”
- Even though games begin with arbitrary or contingent rules, we discover unanticipated features through play and add new rules to make them more interesting—as football added the back-pass and offside rules—so the rule-bound nature of games is not fully transparent at the outset and provides genuine constraints, addressing the worry that obviously constructed rules feel less meaningful.
“we discover things about the games that we play that were unexpected or unanticipated given that initial rule set, and we might find the need to add new rules to the game”
- Work provides four goods beyond money—mastery of a skill set, community and friendship, contribution to society, and social recognition and status—but work is not the only forum through which these can be obtained, so losing work need not mean losing these goods if other fora can provide them.
“they identify four things that people get out of work, apart from an income. They get a sense of mastery over some skill set. They get to form alliances and friendships... They also get to contribute to their societies... they get a sense of social recognition and status”
- Hunter-gatherers were arguably the original affluent leisure society, spending only a couple of hours a day acquiring food and deriving meaning and satisfaction from leisure for the rest, undercutting the assumption that humans have always had to work hard throughout history.
“hunter-gatherers were the original leisure society 'cause they spent very little time every day working. In terms of getting their food, they spent a couple of hours doing that, and then the rest of their day was quite leisurely”
Whether work disappears depends on human choice, not technological inevitability or cyborg competition.
- Humans became evolutionarily successful by occupying the cognitive niche through complex problem-solving, but technology is increasingly ceding that territory to machines, leaving two responses: the cyborg solution of fighting back by becoming more machine-like, or the virtual solution of ceding the niche and retreating to something else.
“technology seems to be pushing humans out of the cognitive niche, it's a concept from evolutionary psychology and evolutionary anthropology that I borrow as a metaphor”
- Whether work fully disappears is not technologically predetermined; it depends on individual, societal, and institutional choices, so Danaher is neither a technological determinist nor a fatalist even while being skeptical that new complementary human tasks will always emerge to absorb mass employment.
“I'm not a technological determinist or a fatalist about these things. I think there are choices that we make individually, societally and institutionally that will make a difference”
- There is a paradox in transhumanism, as Mark O'Connell observed, where the desire to escape the constraints of inefficient biology by replacing it with technology merely substitutes a biological prison for a technological one, since the technological infrastructure you depend on is likely controlled by corporations rather than individuals, potentially making you less free.
“you're replacing the biological prison potentially with a technological prison. So you're substituting one kind of un-freedom for another kind of un-freedom, because who controls the technological infrastructure that you replace yourself with”
- Pursuing cyborg augmentation to stay competitive with machines risks perpetuating the economy in its current form and doubling down on its worst features, since competition for cyborg implants could replace competition for education and worsen inequality if implants are affordable only to an elite few.
“that would probably seem to be a way of perpetuating the economy in its current form... this might double down on many of the worst aspects of the labor market”
Basic income and virtual futures are ethically justified independent of automation's technological trajectory.
- There are strong ethical reasons to favor a basic income that have nothing to do with technology—chiefly that it expands opportunity and freedom of choice in life—and philosophers defended it long before the automation narrative took center stage in the past decade.
“a basic income allows for more opportunity and freedom of choice in life. And that tends to be the main ground on which philosophers have defended the notion. The automation narrative has only really taken center stage in the debate around basic income in the past decade”
- The virtual utopia will still contain power inequalities and scarcity, such as competition to be the top-ranked performer at a game, but if material necessity is removed and everyone has a reasonable quality of life, those power dynamics become less meaningful and salient because people are freer to choose their activities, making it better than the current system rather than perfect.
“One of the reasons why the power dynamics within the existing economy are so meaningful and salient to people is because people have to work to survive and to live. If you cut out that kind of material necessity... they have more freedom then to choose the kinds of activities that they pursue”
Human attachment to work and anxiety about artistic obsolescence reflect concerns that misread human values.
- The jobs likely to resist automation are a poor reassurance because they fall into two unhelpful categories: superstar markets like artistry and entertainment where a handful of elite individuals extract most of the value and which were never jobs for the masses, and human-touch work like elderly care that few people want to perform and that demographic pressures will push toward automation anyway.
“that's never been a job for the masses of people. These are always jobs reserved for the elite few. These are so-called superstar markets, where the handful of individuals tend to extract most of the value from those markets”
- Computer-created art is unlikely to be widely embraced because, as with art forgery, we care about the human origin and story behind an artwork more than the end product itself, so a perfect forgery or algorithmic creation is never valued as highly as a human original.
“there is something in art that we like the human origin and the human story behind it, and that often counts for more than just the end product. And we see this anyway in debates about art and forgery. A perfect forgery is never as good as the original artwork”
- Automation is reaching science itself: researchers at the University of Aberystwyth created robots named Adam and Eve that generate their own hypotheses, experimentally test them, and reach conclusions, demonstrating automation extending into traditionally human cognitive professions.
“the University of Aberystwyth in Wales that have created these robots, Adam and Eve, that run their own experiments. They come up with their own hypotheses and tests their own... experimentally test these hypotheses, reach conclusions”
- Danaher defends ethical behaviorism toward virtual or machine-like others: if they look and act roughly the same as humans in all important respects, then we can have meaningful relationships with them, so robot friends or lovers are possible under certain conditions.
“I define myself as ethical behaviorist when it comes to our interactions with virtual others or machine-like others. In that, if they look and act like humans or are the same roughly as humans in all important respects, then we can have meaningful relationships with them”
Policymakers should map multiple future scenarios against ethical norms before change forces reactive decisions.
- Rather than predicting the future or cheerleading one outcome, the valuable intellectual move is to sketch multiple possible future scenarios and evaluate them against different ethical norms, so we understand the broad axiological landscape into which we are navigating and can act before, not after, dramatic change arrives.
“I try to imagine different possible futures and evaluate them using different ethical norms and principles, so that we get a sense of the broadness of the axiological landscape into which we are navigating”
Human dominance of Earth's biomass has grown from 0.1% ten millennia ago to 98% today.
- Whereas 10,000 years ago humans plus our livestock and pets made up roughly 0.1% of vertebrate biomass on Earth, today that figure is closer to 98%, reflecting unprecedented human dominance of the planet since the agricultural revolution.
“the percentage of the vertebrate biomass on Earth that was either human beings or our controlled animals, our livestock and our pets, was 0.1% of the vertebrate biomass on Earth. And by now it's closer to 98%”