Handing Over Humanity (Ruxandra Teslo)
| Author | Ruxandra Teslo |
|---|---|
| Published | November 11, 2026 |
| Format | Essay |
| Original | Read the original |
Among people building and studying frontier AI, “handoverism” (ceding control of humanity’s future to a benevolent AI) passes for a good outcome; Teslo argues it is undemocratic, practically fragile, historically dangerous, and robs future generations of the ability to remake their world.
Handing Over Humanity (Ruxandra Teslo)
Key claims
- The opening anecdote: an employee at a major AI lab told her, casually, that permanent disempowerment under a hopefully benevolent “AI God” is not his preferred outcome but the only one he can see once superintelligence arrives, and still better than his real fear, that everyone dies.
- She moved to San Francisco three weeks before writing, on what locals describe as the eve of recursive self-improvement, discussed as months away. p(doom) is a constant topic, with many putting it at 10% or higher; a few, like Ginevra Lily Davis in “The Silicon Man,” welcome human succession (“A bootloader, if you will”), and Elon Musk has called spreading superintelligence “a great success condition for a biological bootloader.”
- “Handoverism” is the more common view among those who want humans to survive: humanity survives but entrusts its direction to AI. Scott Alexander’s recent essay favours a fifth, “God” scenario (“please arrange the world as you see fit”); the AI 2040 proposal from prominent safety researchers describes control passing gradually to aligned machines with “a point of no return.” Teslo estimates 30-40% of the AI safety community outside the labs holds some version, higher inside.
- Two camps: enthusiasts, who start from a bleak view of human civilization (Alexander’s 2014 “Meditations on Moloch” ends by suggesting only a benevolent superintelligence can stop competitive spirals), and the resigned, who see themselves as tools of “technocapital” and treat fighting for human control as “loser-coded.” Adjacent ideas: humans as cherished “labradors,” and permanent displacement of labour.
- She had agents survey lab founders’ public statements. Anthropic’s Tom Brown is most explicit (“a handoff where humanity will hand off control to transformative AI”). Dario Amodei’s Machines of Loving Grace borrows from Richard Brautigan’s 1967 poem about being “watched over” by machines, a Rousseauian longing she reads as implying handover, though he elsewhere wants humans to keep “some control.” Sam Altman accepted handover publicly from 2017 but has since said humans should stay in control.
- Greg Brockman, Mustafa Suleyman and Jack Clark reject handoverism, but apart from Suleyman their stated visions don’t square with it: Demis Hassabis and Greg Brockman cite Iain M. Banks’s Culture, where humans are, in Banks’s words, “somewhere between passengers, pets and parasites.”
- Post-work futures are underexplained: work is income but also political and economic leverage, and “live on capital returns” says nothing about how people without capital acquire it. The end-of-disease promise does not obviously require a godlike AI, and we have never accepted any trade-off to speed medicine (no one wants to be Mengele).
- Her objections: a commitment to self-determination rooted in Alexis de Tocqueville’s Democracy in America and his warning of a gentle government that spares people “all the care of thinking and all the trouble of living”; practical fragility, since keeping power responsive took centuries of institutions and feedback, which a one-off alignment cannot replace; the political history of surrendering power for a final vision (Karl Popper’s “Utopia and Violence” and piecemeal reform); and permanence, which freezes 2026’s values and, per Hannah Arendt’s natality, robs descendants of the freedom to begin something new.
Notable quotes
It was said surprisingly casually, the way one states a banal fact about tides.
One scarcely has time to furnish a flat, or, for that matter, to find one.
One common and ostensibly optimistic vision is known as “handover”: humanity cedes control of its future to a godlike AI and we become the comfortably maintained inhabitants of a civilisation whose direction is no longer ours to decide.
Whether or not the belief is correct, it is sincerely held, and it is not “marketing hype.”
Work is more than a source of income: it is also a source of leverage, political as well as economic.
Why should the convictions of 2026 become the permanent constitution of the human future?
Leaving our descendants a civilization frozen in amber, one they cannot reshape by any means sounds to me like robbing them of their birthright.
How it connects
- Bill Gurley amplified the essay in It is critically important that… (tweet), arguing that if Anthropic’s founders believe in handover it belongs in the risk factors of the company’s S-1.
- Dario Amodei, Sam Altman, Elon Musk, Greg Brockman, Demis Hassabis: the lab leaders whose public statements she surveys.
- Alexis de Tocqueville, Karl Popper, Hannah Arendt: the three thinkers behind her case for self-determination.
- Recursive, Self, Improvement (Aashish Reddy, Cosmos Institute): same week, the mechanism-level account of the acceleration she takes as given, ending on the same question of whose control.
- Just Ridiculous — A Famed Philosopher Takes on AI Consciousness (WIRED): Thomas Nagel’s parallel objection, about minds rather than control.
- Workers Are Not Going to Be OK (Azeem Azhar, NYT): the economic version of her point that work is leverage.