Robin Hanson's 'Age of Em': A Brutally Detailed Vision of the Uploaded-Brain Economy

The Most Hansonian Book Ever Written

Robin Hanson — economist at George Mason University, former DARPA and NASA researcher, and one of the most reliably provocative thinkers in the social sciences — has produced a work of futurology so exhaustively detailed, so relentlessly logical, and so cheerfully unsettling that it defies easy categorization. His book on the coming age of "ems" (emulated humans) reads less like science fiction and more like a business school casebook set three generations in the future. The effect is genuinely disorienting: Hanson writes about what swear words uploaded minds might use in the same dry professorial monotone he uses to discuss optimal data-center cooling pipes.

Hanson's intellectual signature — described by economist Bryan Caplan as provoking an immediate "No way! Impossible!" followed by years of genuine rethinking — is on full display throughout. He cites obscure, well-sourced empirical claims (mental fatigue reduces performance by roughly 0.1% per minute; wars are distributed evenly across all possible sizes; experimental painters peak creatively at 46–52 while conceptual artists peak at 24–34) and then methodically rings every possible implication out of them. Whether you agree with his conclusions or not, the machinery of his reasoning is impossible to dismiss.

What Ems Are and Why They Change Everything

Hanson's central premise: artificial general intelligence will remain too difficult to develop in time to reshape the near-future world, but within roughly a century, advances in brain scanning, neuroscience, and computing hardware will enable full neuronal simulation of human minds. Scan a brain at sufficient resolution and simulate it neuron by neuron, and you get an emulated person — an "em" — that processes information the way the original brain does, with all the same memories, personality, and cognitive quirks.

Ems differ from biological humans in three transformative ways. First, they require no physical body — no food, no water, no medicine. In virtual environments, any luxury is computationally conjurable at near-zero cost. Second, they can run faster or slower than biological time. A sufficiently powerful hardware cluster could let an em subjectively experience a century in one objective week; a hardware-starved em could slow its processing so that it experiences one subjective week every twenty objective years. Third — and most economically disruptive — ems can be copied. One uploaded mind plus spare computing capacity equals a thousand identical minds, each capable of independent work, conversation, romance, or combat.

The Psychopathic Logic of the Em Economy

The copy-ability of ems detonates conventional labor economics. If a company needs a thousand Ruby programmers, it doesn't recruit a thousand workers — it finds one exceptionally talented person, copies her a thousand times, and pays each copy at subsistence level, which Hanson defines as the bare minimum compute cost to keep an em running. With unlimited labor supply, wages converge toward that floor. The majority of all ems in this future will exist at such subsistence wages.

But the workarounds are where the book becomes genuinely strange. Hanson's solutions to labor-management problems have a quality one reviewer compared to the logic of a psychopath solving a puzzle — technically correct, completely alien to ordinary moral intuitions, and oddly compelling. Consider the sleep problem: a company that wants a surgeon working continuously can simply let one em-copy sleep a full night, then paste that rested copy at the start of each surgery, delete it when it tires, and paste the rested version again. The em never accumulates fatigue because it is repeatedly reset.

Vacation works similarly: give one "trunk" em a genuine holiday, copy that brain-state a thousand times each morning, work all copies for twenty-four hours, then delete them. Every working em subjectively remembers a life of constant leisure and performs accordingly, while from the company's ledger, 99.9% of ems are productive at any moment. Retirement savings can be pooled across thousands of copies of the same person, with all but one randomly selected copy voluntarily ceasing to exist at retirement age, leaving that survivor with a nest egg thousands of times what any single copy could have accumulated alone.

Copy Clans, Speed Differentials, and the Social Fabric of Uploaded Life

Social life in the em era is projected to be no less bizarre. Because companies will only fund copying of the most economically valuable minds, the trillions of ems populating this future will mostly be copies of a relatively small number of exceptionally talented originals. Hanson introduces the concept of "copy clans" — governance and social structures organized around all living copies of the same original person. These clans become natural units for finance, legal liability, and political representation.

Romantic and social relationships acquire new layers of strangeness. Any given copy might consult statistical records of how prior interactions between copies of their own template and copies of a prospective partner's template have gone historically — an 87% receptivity rate, say — before initiating contact. Relationship advice flows from the collective institutional memory of the copy clan. Speed differentials create a novel form of relationship breakdown: if one partner's employer runs simulations at 10,000 times human speed and the other's runs at 100 times, the resulting subjective time gap makes ordinary interaction impossible. The proposed fix — one partner purchasing the other's source code and instantiating a private copy to socialize with — is presented with the same matter-of-fact calm as any other logistical arrangement.

The Book as Manifesto, Mind-Expander, and Cautionary Example

Hanson is upfront about his uncertainty: he estimates that conditional on his core assumptions holding, at least 30% of his specific predictions will prove usefully accurate, and unconditionally perhaps 10%. That epistemic humility sits somewhat awkwardly alongside the book's two pages speculating on future profanity, but Hanson's broader intellectual project justifies the exercise in multiple ways. The book functions simultaneously as a primer on recurring themes in Hanson's thought — near versus far thinking modes, signaling theory, inside versus outside views, the farmer-forager dichotomy — delivered through dozens of applied examples rather than abstract exposition. It also functions as a manifesto for institutions Hanson has long advocated, particularly prediction markets, which he invented and has spent decades promoting. Embedded in a futurological scenario, their advantages reach readers who might otherwise never encounter the idea.

The futurology itself carries a genuine cognitive benefit independent of its accuracy. Most people's visions of the future are timid — aliens with green skin instead of genuinely alien minds. Hanson's scenario forces serious engagement with the possibility that descendants could be simultaneously incomprehensibly wealthy and subjectively impoverished; shorter-lived in some senses and immortal in others; less likely to travel through space; less progressive; operating on timescales from fractions of seconds to geological epochs. Whether or not ems ever exist, the exercise of modeling their world rigorously challenges assumptions about what aspects of human life are contingent and what are truly fundamental.