superintelligence
Hypothesized AI vastly exceeding human intelligence across all domains.
As defined in cwt_alison-gopnik
Other definitions
An AI system significantly smarter than all humans; the key inflection point in Eliezer's doom scenario where AI systems become impossible to control or contain
Live: Eliezer Yudkowsky - Is Artificial General Intelligence too Dangerous to Build?
Where it comes up · 52
- Robot Plumbers, Robot Armies, and Our Imminent A.I. Future | Interesting Times with Ross Douthat37 claims
- AI 2027: month-by-month model of intelligence explosion — Scott Alexander & Daniel Kokotajlo24 claims
- Nick Bostrom: What Happens When AI Evolves Faster Than Humans?15 claims
- The AI Revolution Is Underhyped | Eric Schmidt | TED12 claims
- 420 - Countdown To Superintelligence10 claims
- Leopold Aschenbrenner — 2027 AGI, China/US super-intelligence race, & the return of history9 claims
- Live: Eliezer Yudkowsky - Is Artificial General Intelligence too Dangerous to Build?8 claims
- Ilya Sutskever – We're moving from the age of scaling to the age of research8 claims
and 44 more
Most important claims · 10 of 236Search all claims for “superintelligence” →
- The nuclear-weapons analogy for AI fails because AI is not a self-contained pure weapon but is closer to industrialization itself—a general-purpose economic transformation with thousands of applications—so applying the 'government may seize the monopoly builder' logic would imply the government could have requisitioned any factory during the Industrial Revolution, which free societies rightly did not do.Dwarkesh Patel Podcast - The most important question nobody's asking about AI
- The worry about AGI is not that robot overlords spontaneously become evil; it is a story of what mismatches in competence and power produce in the absence of perfect alignment—just as humans, pursuing goals ants and farm animals cannot glimpse, treat them in ways totally inscrutable and destructive to them, a superintelligence not perfectly aligned with human well-being could spontaneously treat us according to goals we can no longer contemplate, without any hostility being required.312 - The Trouble With Ai
- Progress is a pretty smooth exponential rather than a sudden jump, so the month after superintelligence arrives would bring little novel change, since human progress has long compounded and each decade is more different than the last, and people overestimate the significance of the moment.Sam Altman - How to Make an Abundant Future - [Invest Like the Best, EP.484]
- A single all-powerful superintelligence (a 'John von Neumann times 10') is less likely to be the locus of big impacts; the key advantage of AI lies in the scope of social learning a society of AIs can do, because humans achieved their feats not through individual brainpower but through the power of cultural evolution — a network of minds gnawing at problems over long periods, accumulating both tools and the cognitive tools to address them.Joseph Henrich — Humans defeated smarter species with cultural evolution
- Superintelligence is any intellect that radically outperforms humanity across all practically relevant fields, including scientific creativity, social skills, and general wisdom.Nick Bostrom on Superintelligence
- The fastest historical precedent for converting manufacturing capacity during wartime was WWII bomber factory conversion, which took approximately 3 years from decision to achieve the production rate of 1 bomber per hour; the AI 2027 scenario estimates superintelligence-driven robot factory conversion could achieve comparable results in approximately 1 year due to elimination of human bureaucratic delays and government wartime urgency.AI 2027: month-by-month model of intelligence explosion — Scott Alexander & Daniel Kokotajlo
- Humans are not naturally equipped to deal with things smarter than themselves; we're used to being the apex intelligence, so facing superintelligence is something humans have never adapted to.Geoffrey Hinton - Two Paths to Intelligence
- The critical inflection point is when AI can fully substitute for a machine-learning researcher, because at that moment the world effectively gains hundreds of millions of AI researchers working on the next generation plus millions of AI scientists on other problems, producing a discontinuous leap in capability—an intelligence explosion—plausibly arriving in the early 2030s with large error bars.467 - Ea Ai And The End Of Work
- Humanity faces a high-wire act: building superintelligence misaligned would be catastrophic, but failing ever to develop superintelligence would itself plausibly constitute an existential catastrophe, because some of our existential problems may only be solvable in the presence of greater intelligence.385 - Ai Utopia
- Despite existential risks, superintelligence development should be pursued because all possible paths to truly great futures for humanity lead through the portal of superintelligence development.Nick Bostrom: What Happens When AI Evolves Faster Than Humans?