{"meta":{"updatedAt":"2026-11-06"},"index":{"byClaim":[["pr-a-turing-1950-imitation-game",["predictions:predictions-on-record"]],["pr-a-turing-1950-end-of-century",["predictions:predictions-on-record"]],["pr-a-dartmouth-1955-summer",["predictions:predictions-on-record"]],["pr-a-simon-newell-1958-chess",["predictions:predictions-on-record"]],["pr-a-good-1965-ultraintelligent-machine",["predictions:predictions-on-record"]],["pr-a-michie-1973-opinion-poll",["predictions:predictions-on-record"]],["pr-a-lighthill-1973-correspondents-1980s",["predictions:predictions-on-record"]],["pr-a-lighthill-1973-correspondents-2000",["predictions:predictions-on-record"]],["pr-b-degaris-1989-commonplace",["predictions:predictions-on-record"]],["pr-b-vinge-1993-thirty-years",["predictions:predictions-on-record"]],["pr-b-vinge-1993-2005-2030",["predictions:predictions-on-record"]],["pr-b-yudkowsky-2000-2035-upper-bound",["predictions:predictions-on-record"]],["pr-b-bostrom-1998-first-third",["predictions:predictions-on-record"]],["pr-b-bostrom-1998-hardware-2004-2008",["predictions:predictions-on-record"]],["pr-b-moravec-1998-hardware-2020s",["predictions:predictions-on-record"]],["pr-b-moravec-1999-before-2050",["predictions:predictions-on-record"]],["pr-b-kurzweil-2001-brain-capacity",["predictions:predictions-on-record"]],["pr-b-kurzweil-2001-1000-dollars-2023",["predictions:predictions-on-record"]],["pr-b-legg-2011-human-level",["predictions:predictions-on-record"]],["pr-c-baum-2011-agi09",["predictions:predictions-on-record"]],["pr-c-muller-bostrom-2014-hlmi",["predictions:predictions-on-record"]],["pr-c-muller-bostrom-2014-superintelligence",["predictions:predictions-on-record"]],["pr-c-grace-2017-all-tasks",["predictions:predictions-on-record"]],["pr-c-grace-2017-all-jobs",["predictions:predictions-on-record"]],["pr-c-grace-2017-ten-years",["predictions:predictions-on-record"]],["pr-c-zhang-2022-hlmi-2060",["predictions:predictions-on-record"]],["pr-c-espai-2022-hlmi-2059",["predictions:predictions-on-record"]],["pr-c-grace-2024-every-task",["predictions:predictions-on-record"]],["pr-c-grace-2024-occupations",["predictions:predictions-on-record"]],["pr-d-longbets-1-turing-test-2029",["predictions:predictions-on-record"]],["pr-e-brooks-2018-next-big-thing",["predictions:predictions-on-record"]],["pr-e-brooks-2018-mouse",["predictions:predictions-on-record"]],["pr-e-brooks-2018-dog",["predictions:predictions-on-record"]],["pr-e-karnofsky-2021-transformative-ai",["predictions:predictions-on-record"]],["pr-e-carlsmith-2022-catastrophe-2070",["predictions:predictions-on-record"]],["pr-e-cotra-2020-bio-anchors-as-reported-2022",["predictions:predictions-on-record"]],["pr-e-cotra-2022-update",["predictions:predictions-on-record"]],["pr-e-aschenbrenner-2024-agi-2027",["predictions:predictions-on-record"]],["pr-e-bengio-et-al-2023-this-decade-or-next",["predictions:predictions-on-record"]],["pr-f-openai-2023-governance-ten-years",["predictions:predictions-on-record"]],["pr-f-openai-2023-superalignment-this-decade",["predictions:predictions-on-record"]],["pr-f-altman-2024-few-thousand-days",["predictions:predictions-on-record"]],["pr-f-amodei-2024-powerful-ai-2026",["predictions:predictions-on-record"]],["pr-f-altman-2025-agents-join-workforce",["predictions:predictions-on-record"]],["pr-f-altman-2025-2035-capacity",["predictions:predictions-on-record"]],["pr-f-anthropic-2025-late-2026-early-2027",["predictions:predictions-on-record"]],["pr-f-ai2027-2025-next-decade",["predictions:predictions-on-record"]],["pr-f-google-deepmind-2025-by-2030",["predictions:predictions-on-record"]],["pr-f-amodei-2025-country-of-geniuses",["predictions:predictions-on-record"]],["pr-f-altman-2025-novel-insights-2026",["predictions:predictions-on-record"]],["pr-f-altman-2025-robots-2027",["predictions:predictions-on-record"]],["pr-f-amodei-2026-one-to-two-years",["predictions:predictions-on-record"]],["pr-f-amodei-2026-few-years-everything",["predictions:predictions-on-record"]],["pr-a-simon-newell-1958-theorem",["predictions:predictions-on-record"]],["pr-a-simon-newell-1958-music",["predictions:predictions-on-record"]],["pr-a-simon-newell-1958-psychology",["predictions:predictions-on-record"]],["pr-b-vinge-1983-twenty-to-hundred-years",["predictions:predictions-on-record"]],["pr-b-vinge-1983-next-twenty-years",["predictions:predictions-on-record"]],["pr-e-legg-2009-mode-2025",["predictions:predictions-on-record"]],["pr-e-legg-2010-log-normal",["predictions:predictions-on-record"]],["pr-e-legg-2011-log-normal",["predictions:predictions-on-record"]],["pr-c-fhi-2011-hlmi",["predictions:predictions-on-record"]],["pr-d-metaculus-5121-capture-2025-01-05",["predictions:predictions-on-record"]],["pr-f-openai-2025-very-small-discoveries",["predictions:predictions-on-record"]],["pr-f-openai-2025-more-significant-discoveries",["predictions:predictions-on-record"]]],"marks":[["good-1965",[{"collectionId":"superintelligence","propositionId":"intelligence-explosion","claimId":"ts-a1-good-1965","text":"Earliest located in Propositions about superintelligence, 1863–2025 for A1 Intelligence explosion"},{"collectionId":"superintelligence","propositionId":"intelligence-explosion","claimId":"ts-a1-good-1965","text":"Earliest located use of the name “intelligence explosion” in Propositions about superintelligence, 1863–2025 for A1 Intelligence explosion"},{"collectionId":"superintelligence","propositionId":"ai-automated-ai-research","claimId":"ts-a5-good-1965","text":"Earliest located in Propositions about superintelligence, 1863–2025 for A5 AI-automated AI research"}]],["turing-1950-computing-machinery",[{"collectionId":"antecedent-texts","propositionId":"machines-only-do-what-they-are-ordered","claimId":"pre-u01-turing-1950-contests","text":"Earliest located use of the name “Lady Lovelace's objection” in Antecedent texts: machine intelligence before 1965 for U01 A machine can only do what it is ordered to do"},{"collectionId":"superintelligence","propositionId":"designed-deception-of-testers","claimId":"ts-c2-turing-1950","text":"Earliest located in Propositions about superintelligence, 1863–2025 for C2 Systems designed to mislead their testers"}]],["vinge-1983-first-word",[{"collectionId":"antecedent-texts","propositionId":"technological-singularity","claimId":"pre-u03-vinge-1983-first-word","text":"Earliest located in Antecedent texts: machine intelligence before 1965 for U03 Machine intelligence will bring a singularity"}]],["vinge-1993-coming-technological-singularity",[{"collectionId":"superintelligence","propositionId":"fast-or-hard-takeoff","claimId":"ts-a4-vinge-1993","text":"Earliest located in Propositions about superintelligence, 1863–2025 for A4 Fast or hard takeoff"},{"collectionId":"superintelligence","propositionId":"race-compression","claimId":"ts-a6-vinge-1993","text":"Earliest located in Propositions about superintelligence, 1863–2025 for A6 Strategic or race compression"},{"collectionId":"superintelligence","propositionId":"strategic-concealment","claimId":"ts-b2-vinge-1993","text":"Earliest located in Propositions about superintelligence, 1863–2025 for B2 Strategic concealment"},{"collectionId":"superintelligence","propositionId":"competition-defeats-restraint","claimId":"ts-c3-vinge-1993","text":"Earliest located in Propositions about superintelligence, 1863–2025 for C3 Competition defeats restraint"},{"collectionId":"risk-debate","propositionId":"fast-or-hard-takeoff","claimId":"ts-a4-vinge-1993","text":"Earliest located in Arguments about catastrophic risk from advanced AI, 1863–2026 for A4 Fast or hard takeoff"},{"collectionId":"risk-debate","propositionId":"control-by-design","claimId":"rd-r10b-vinge-1993-coming-technological-singularity","text":"Earliest located text contesting R10b Advanced AI can be kept under control by design in Arguments about catastrophic risk from advanced AI, 1863–2026"},{"collectionId":"risk-debate","propositionId":"incentives-produce-safety","claimId":"rd-r17-vinge-1993-coming-technological-singularity","text":"Earliest located text contesting R17 Economic incentives will produce safety in Arguments about catastrophic risk from advanced AI, 1863–2026"}]],["staring-into-the-singularity",[{"collectionId":"yudkowsky-arguments","propositionId":"agi-decades-away","claimId":"yk-y13-staring-into-the-singularity-c","text":"Earliest located text contesting Y13 AGI is decades away, so the problem can wait in Eliezer Yudkowsky’s arguments on common claims about AI, 1996–2026"},{"collectionId":"yudkowsky-arguments","propositionId":"build-superintelligence-fast","claimId":"yk-y22-staring-into-the-singularity-a","text":"Earliest located in Eliezer Yudkowsky’s arguments on common claims about AI, 1996–2026 for Y22 Build superintelligence as fast as possible"},{"collectionId":"machine-succession","propositionId":"vast-superiority","claimId":"ht-w05-yudkowsky-2000-staring-into-the-singularity","text":"Earliest located in Machine succession, scale and acceleration: propositions in located texts, 1863–2026 for W05 Superiority by many orders of magnitude"},{"collectionId":"machine-succession","propositionId":"accelerate-machine-intelligence","claimId":"ht-w08-yudkowsky-1996-staring-into-the-singularity","text":"Earliest located in Machine succession, scale and acceleration: propositions in located texts, 1863–2026 for W08 Accelerate rather than restrain"}]],["carlsmith-2022-power-seeking",[{"collectionId":"risk-debate","propositionId":"loss-of-control-route","claimId":"rd-r08-carlsmith-2022-power-seeking","text":"Earliest located use of the name “disempowerment” in Arguments about catastrophic risk from advanced AI, 1863–2026 for R08 Existential risk from misalignment or loss of control"}]],["degaris-1989-what-if-ai-succeeds",[{"collectionId":"machine-succession","propositionId":"species-dominance-conflict","claimId":"ht-w04-degaris-1989-what-if-ai-succeeds","text":"Earliest located in Machine succession, scale and acceleration: propositions in located texts, 1863–2026 for W04 Conflict over building superhuman machines"},{"collectionId":"machine-succession","propositionId":"species-dominance-conflict","claimId":"ht-w04-degaris-1989-what-if-ai-succeeds","text":"Earliest located use of the name “Cosmist” in Machine succession, scale and acceleration: propositions in located texts, 1863–2026 for W04 Conflict over building superhuman machines"}]]],"chronicle":{"years":[{"year":1637,"topics":{"ideas":2},"stated":0},{"year":1843,"topics":{"ideas":1},"stated":0},{"year":1863,"topics":{"ideas":10},"stated":0},{"year":1865,"topics":{"ideas":3},"stated":0},{"year":1872,"topics":{"ideas":12},"stated":0},{"year":1923,"topics":{"ideas":3},"stated":0},{"year":1943,"topics":{"ideas":1},"stated":0},{"year":1947,"topics":{"ideas":4},"stated":0},{"year":1948,"topics":{"ideas":7},"stated":0},{"year":1949,"topics":{"ideas":4},"stated":0},{"year":1950,"topics":{"ideas":13},"stated":0},{"year":1951,"topics":{"ideas":14},"stated":0},{"year":1952,"topics":{"ideas":2},"stated":0},{"year":1955,"topics":{"ideas":32},"stated":0},{"year":1956,"topics":{"ideas":1},"stated":0},{"year":1957,"topics":{"ideas":1},"stated":0},{"year":1958,"topics":{"ideas":8},"stated":0},{"year":1959,"topics":{"ideas":7},"stated":0},{"year":1960,"topics":{"ideas":12},"stated":0},{"year":1961,"topics":{"ideas":14},"stated":0},{"year":1965,"topics":{"ideas":9},"stated":0},{"year":1966,"topics":{"ideas":21},"stated":0},{"year":1968,"topics":{"ideas":5},"stated":0},{"year":1969,"topics":{"ideas":2},"stated":0},{"year":1971,"topics":{"ideas":4},"stated":0},{"year":1972,"topics":{"ideas":1},"stated":0},{"year":1973,"topics":{"ideas":5},"stated":0},{"year":1974,"topics":{"ideas":2},"stated":0},{"year":1975,"topics":{"ideas":1},"stated":0},{"year":1976,"topics":{"ideas":1},"stated":0},{"year":1977,"topics":{"people":2,"ideas":8},"stated":0},{"year":1979,"topics":{"ideas":10},"stated":0},{"year":1980,"topics":{"ideas":2},"stated":0},{"year":1981,"topics":{"people":2},"stated":0},{"year":1982,"topics":{"ideas":4},"stated":0},{"year":1983,"topics":{"ideas":7},"stated":0},{"year":1984,"topics":{"ideas":21},"stated":0},{"year":1985,"topics":{"ideas":9},"stated":0},{"year":1986,"topics":{"ideas":18},"stated":0},{"year":1987,"topics":{"ideas":1},"stated":0},{"year":1988,"topics":{"ideas":6},"stated":0},{"year":1989,"topics":{"ideas":26},"stated":0},{"year":1990,"topics":{"ideas":9},"stated":0},{"year":1991,"topics":{"ideas":14},"stated":0},{"year":1992,"topics":{"ideas":15},"stated":0},{"year":1993,"topics":{"ideas":27,"people":3},"stated":0},{"year":1994,"topics":{"ideas":18},"stated":0},{"year":1995,"topics":{"ideas":10,"people":1},"stated":0},{"year":1996,"topics":{"ideas":83},"stated":0},{"year":1997,"topics":{"ideas":19},"stated":0},{"year":1998,"topics":{"ideas":56},"stated":0},{"year":1999,"topics":{"ideas":33,"people":1,"institutions":1},"stated":0},{"year":2000,"topics":{"ideas":77},"stated":0},{"year":2001,"topics":{"ideas":55},"stated":0},{"year":2002,"topics":{"ideas":32},"stated":0},{"year":2003,"topics":{"ideas":9},"stated":0},{"year":2004,"topics":{"ideas":14,"people":3},"stated":0},{"year":2005,"topics":{"ideas":20,"people":1},"stated":0},{"year":2006,"topics":{"ideas":26,"people":1},"stated":0},{"year":2007,"topics":{"ideas":21},"stated":0},{"year":2008,"topics":{"ideas":67,"people":3,"institutions":1},"stated":0},{"year":2009,"topics":{"ideas":13,"people":1},"stated":0},{"year":2010,"topics":{"people":3,"ideas":14,"institutions":1,"governance":4},"stated":0},{"year":2011,"topics":{"people":3,"ideas":20,"institutions":2},"stated":0},{"year":2012,"topics":{"people":4,"ideas":20,"institutions":2},"stated":0},{"year":2013,"topics":{"ideas":47,"people":4,"institutions":2,"capital":1},"stated":0},{"year":2014,"topics":{"institutions":2,"ideas":62,"people":5},"stated":0},{"year":2015,"topics":{"people":23,"ideas":133,"institutions":3,"capital":1},"stated":0},{"year":2016,"topics":{"ideas":88,"people":2},"stated":0},{"year":2017,"topics":{"people":16,"institutions":2,"ideas":133,"governance":2},"stated":0},{"year":2018,"topics":{"people":9,"ideas":163,"institutions":8,"governance":1},"stated":0},{"year":2019,"topics":{"capital":1,"compute":1,"people":11,"governance":19,"ideas":176,"institutions":4},"stated":0},{"year":2020,"topics":{"people":6,"ideas":98,"institutions":1,"governance":2},"stated":0},{"year":2021,"topics":{"people":27,"ideas":178,"compute":1,"institutions":23,"governance":4,"capital":6},"stated":0},{"year":2022,"topics":{"people":15,"ideas":233,"institutions":6,"governance":3,"compute":2,"capital":1},"stated":0},{"year":2023,"topics":{"capital":14,"compute":7,"institutions":25,"people":55,"governance":239,"ideas":414},"stated":0},{"year":2024,"topics":{"capital":31,"compute":19,"governance":367,"people":103,"ideas":185,"institutions":32},"stated":1},{"year":2025,"topics":{"governance":795,"compute":69,"capital":61,"people":145,"institutions":26,"ideas":191},"stated":7},{"year":2026,"topics":{"governance":619,"capital":30,"compute":32,"institutions":23,"ideas":99,"people":139},"stated":19},{"year":2027,"topics":{},"stated":8},{"year":2028,"topics":{},"stated":6},{"year":2029,"topics":{},"stated":3},{"year":2030,"topics":{},"stated":3},{"year":2032,"topics":{},"stated":2},{"year":2035,"topics":{},"stated":2},{"year":2039,"topics":{},"stated":1},{"year":2042,"topics":{},"stated":1},{"year":2050,"topics":{},"stated":1}],"scheduled":[],"statedBy":[]},"collections":[{"target":{"kind":"propositions","id":"antecedent-texts"},"title":"Antecedent texts: machine intelligence before 1965","claimIds":[],"records":[]},{"target":{"kind":"propositions","id":"superintelligence"},"title":"Propositions about superintelligence, 1863–2025","claimIds":[],"records":[]},{"target":{"kind":"propositions","id":"risk-debate"},"title":"Arguments about catastrophic risk from advanced AI, 1863–2026","claimIds":[],"records":[]},{"target":{"kind":"propositions","id":"yudkowsky-arguments"},"title":"Eliezer Yudkowsky’s arguments on common claims about AI, 1996–2026","claimIds":[],"records":[]},{"target":{"kind":"propositions","id":"machine-succession"},"title":"Machine succession, scale and acceleration: propositions in located texts, 1863–2026","claimIds":[],"records":[]},{"target":{"kind":"propositions","id":"economics-of-machine-intelligence"},"title":"Economics of machine intelligence: texts from 1952 to 2026 that state, contest or bear on its propositions","claimIds":[],"records":[]},{"target":{"kind":"propositions","id":"orbital-compute-arguments"},"title":"Arguments about compute in orbit","claimIds":[],"records":[]},{"target":{"kind":"propositions","id":"ai-bio-arguments"},"title":"Arguments about AI and biological risk","claimIds":[],"records":[]},{"target":{"kind":"propositions","id":"ai-in-science"},"title":"AI in science","claimIds":[],"records":[]},{"target":{"kind":"documents"},"title":"Frontier safety frameworks","claimIds":[]},{"target":{"kind":"compare","id":"frontier-safety-frameworks"},"title":"Compare frameworks","claimIds":[]},{"target":{"kind":"control","id":"openai"},"title":"OpenAI’s control structure, 2015–2026","claimIds":[]},{"target":{"kind":"control","id":"anthropic"},"title":"Anthropic’s control structure, 2021–2026","claimIds":[]},{"target":{"kind":"commitments"},"title":"External commitments and obligations","claimIds":[]},{"target":{"kind":"question","id":"frameworks-to-obligations"},"title":"Frameworks, commitments and legal obligations, 2023–2026","claimIds":[]},{"target":{"kind":"reading"},"title":"Founding and early leadership","claimIds":[]}]},"version":"lq5ag3","build":"index-BTwlRauQ.js","packed":1,"at":{"entities":[24,71,163,3,6,13,14,1270,14,199,6,9,16,131,17,26,194,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1],"claims":[6599,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1],"sources":[],"evidence":[]},"records":{"entities":{"n":46,"fields":{"id":["good-1965","superalignment-announcement","turing-1950-computing-machinery","dartmouth-proposal-1955","vinge-1983-first-word","vinge-1993-coming-technological-singularity","staring-into-the-singularity","th-bostrom-1998-how-long-before-superintelligence","th-lighthill-1973-general-survey","altman-2024-intelligence-age","aschenbrenner-2024-situational-awareness","bengio-2023-managing-ai-risks","carlsmith-2022-power-seeking","degaris-1989-what-if-ai-succeeds","kurzweil-2001-law-of-accelerating-returns","moravec-1999-rise-of-the-robots","simon-newell-1958-heuristic-problem-solving","michie-1973-machines-and-theory-of-intelligence","moravec-1998-when-will-hardware-match-brain","long-bets-1-turing-test-2029","legg-2011-qa-on-risks-from-ai","baum-goertzel-goertzel-2011-how-long-until-human-level-ai","muller-bostrom-2014-future-progress-in-ai","grace-et-al-2017-when-will-ai-exceed-human-performance","zhang-et-al-2022-forecasting-ai-progress","ai-impacts-2022-expert-survey-on-progress-in-ai","grace-et-al-2024-thousands-of-ai-authors","brooks-2018-my-dated-predictions","karnofsky-2021-ai-timelines","cotra-2022-two-year-update","amodei-2024-machines-of-loving-grace","altman-2025-reflections","altman-2025-three-observations","anthropic-2025-recommendations-to-ostp","ai-2027-scenario","google-deepmind-2025-technical-agi-safety","amodei-2025-urgency-of-interpretability","altman-2025-gentle-singularity","amodei-2026-adolescence-of-technology","openai-2023-governance-of-superintelligence","legg-2009-tick-tock-bing","legg-2010-goodbye-2010","legg-2011-goodbye-2011","sandberg-bostrom-2011-machine-intelligence-survey","openai-2025-ai-progress-and-recommendations","metaculus-5121-capture-2025-01-05"],"name":["Speculations Concerning the First Ultraintelligent Machine","Introducing Superalignment","Computing Machinery and Intelligence","A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence","First Word","The Coming Technological Singularity: How to Survive in the Post-Human Era","Staring into the Singularity","How Long Before Superintelligence?","Artificial Intelligence: A General Survey","The Intelligence Age","Situational Awareness: The Decade Ahead","Managing AI Risks in an Era of Rapid Progress","Is Power-Seeking AI an Existential Risk?","What If AI Succeeds? The Rise of the Twenty-First Century Artilect","The Law of Accelerating Returns","Rise of the Robots","Heuristic Problem Solving: The Next Advance in Operations Research","Machines and the theory of intelligence","When will computer hardware match the human brain?","Long Bet 1: By 2029 no computer - or \"machine intelligence\" - will have passed the Turing Test","Q&A with Shane Legg on risks from AI","How long until human-level AI? Results from an expert assessment","Future Progress in Artificial Intelligence: A Survey of Expert Opinion","When Will AI Exceed Human Performance? Evidence from AI Experts","Forecasting AI Progress: Evidence from a Survey of Machine Learning Researchers","2022 Expert Survey on Progress in AI","Thousands of AI Authors on the Future of AI","My Dated Predictions","AI Timelines: Where the Arguments, and the \"Experts,\" Stand","Two-year update on my personal AI timelines","Machines of Loving Grace","Reflections","Three Observations","Anthropic's Recommendations to OSTP for the U.S. AI Action Plan","AI 2027","An Approach to Technical AGI Safety and Security","The Urgency of Interpretability","The Gentle Singularity","The Adolescence of Technology","Governance of superintelligence","Tick, tock, tick, tock… BING","Goodbye 2010","Goodbye 2011, hello 2012","Machine Intelligence Survey","AI progress and recommendations","When will the first general AI system be devised, tested, and publicly announced? (Metaculus question 5121, as captured on 2025-01-05)"],"type":["document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document","document"]}},"claims":{"n":65,"fields":{"id":["pr-a-turing-1950-imitation-game","pr-a-turing-1950-end-of-century","pr-a-dartmouth-1955-summer","pr-a-simon-newell-1958-chess","pr-a-good-1965-ultraintelligent-machine","pr-a-michie-1973-opinion-poll","pr-a-lighthill-1973-correspondents-1980s","pr-a-lighthill-1973-correspondents-2000","pr-b-degaris-1989-commonplace","pr-b-vinge-1993-thirty-years","pr-b-vinge-1993-2005-2030","pr-b-yudkowsky-2000-2035-upper-bound","pr-b-bostrom-1998-first-third","pr-b-bostrom-1998-hardware-2004-2008","pr-b-moravec-1998-hardware-2020s","pr-b-moravec-1999-before-2050","pr-b-kurzweil-2001-brain-capacity","pr-b-kurzweil-2001-1000-dollars-2023","pr-b-legg-2011-human-level","pr-c-baum-2011-agi09","pr-c-muller-bostrom-2014-hlmi","pr-c-muller-bostrom-2014-superintelligence","pr-c-grace-2017-all-tasks","pr-c-grace-2017-all-jobs","pr-c-grace-2017-ten-years","pr-c-zhang-2022-hlmi-2060","pr-c-espai-2022-hlmi-2059","pr-c-grace-2024-every-task","pr-c-grace-2024-occupations","pr-d-longbets-1-turing-test-2029","pr-e-brooks-2018-next-big-thing","pr-e-brooks-2018-mouse","pr-e-brooks-2018-dog","pr-e-karnofsky-2021-transformative-ai","pr-e-carlsmith-2022-catastrophe-2070","pr-e-cotra-2020-bio-anchors-as-reported-2022","pr-e-cotra-2022-update","pr-e-aschenbrenner-2024-agi-2027","pr-e-bengio-et-al-2023-this-decade-or-next","pr-f-openai-2023-governance-ten-years","pr-f-openai-2023-superalignment-this-decade","pr-f-altman-2024-few-thousand-days","pr-f-amodei-2024-powerful-ai-2026","pr-f-altman-2025-agents-join-workforce","pr-f-altman-2025-2035-capacity","pr-f-anthropic-2025-late-2026-early-2027","pr-f-ai2027-2025-next-decade","pr-f-google-deepmind-2025-by-2030","pr-f-amodei-2025-country-of-geniuses","pr-f-altman-2025-novel-insights-2026","pr-f-altman-2025-robots-2027","pr-f-amodei-2026-one-to-two-years","pr-f-amodei-2026-few-years-everything","pr-a-simon-newell-1958-theorem","pr-a-simon-newell-1958-music","pr-a-simon-newell-1958-psychology","pr-b-vinge-1983-twenty-to-hundred-years","pr-b-vinge-1983-next-twenty-years","pr-e-legg-2009-mode-2025","pr-e-legg-2010-log-normal","pr-e-legg-2011-log-normal","pr-c-fhi-2011-hlmi","pr-d-metaculus-5121-capture-2025-01-05","pr-f-openai-2025-very-small-discoveries","pr-f-openai-2025-more-significant-discoveries"],"subject":["turing-1950-computing-machinery","turing-1950-computing-machinery","dartmouth-proposal-1955","simon-newell-1958-heuristic-problem-solving","good-1965","michie-1973-machines-and-theory-of-intelligence","th-lighthill-1973-general-survey","th-lighthill-1973-general-survey","degaris-1989-what-if-ai-succeeds","vinge-1993-coming-technological-singularity","vinge-1993-coming-technological-singularity","staring-into-the-singularity","th-bostrom-1998-how-long-before-superintelligence","th-bostrom-1998-how-long-before-superintelligence","moravec-1998-when-will-hardware-match-brain","moravec-1999-rise-of-the-robots","kurzweil-2001-law-of-accelerating-returns","kurzweil-2001-law-of-accelerating-returns","legg-2011-qa-on-risks-from-ai","baum-goertzel-goertzel-2011-how-long-until-human-level-ai","muller-bostrom-2014-future-progress-in-ai","muller-bostrom-2014-future-progress-in-ai","grace-et-al-2017-when-will-ai-exceed-human-performance","grace-et-al-2017-when-will-ai-exceed-human-performance","grace-et-al-2017-when-will-ai-exceed-human-performance","zhang-et-al-2022-forecasting-ai-progress","ai-impacts-2022-expert-survey-on-progress-in-ai","grace-et-al-2024-thousands-of-ai-authors","grace-et-al-2024-thousands-of-ai-authors","long-bets-1-turing-test-2029","brooks-2018-my-dated-predictions","brooks-2018-my-dated-predictions","brooks-2018-my-dated-predictions","karnofsky-2021-ai-timelines","carlsmith-2022-power-seeking","cotra-2022-two-year-update","cotra-2022-two-year-update","aschenbrenner-2024-situational-awareness","bengio-2023-managing-ai-risks","openai-2023-governance-of-superintelligence","superalignment-announcement","altman-2024-intelligence-age","amodei-2024-machines-of-loving-grace","altman-2025-reflections","altman-2025-three-observations","anthropic-2025-recommendations-to-ostp","ai-2027-scenario","google-deepmind-2025-technical-agi-safety","amodei-2025-urgency-of-interpretability","altman-2025-gentle-singularity","altman-2025-gentle-singularity","amodei-2026-adolescence-of-technology","amodei-2026-adolescence-of-technology","simon-newell-1958-heuristic-problem-solving","simon-newell-1958-heuristic-problem-solving","simon-newell-1958-heuristic-problem-solving","vinge-1983-first-word","vinge-1983-first-word","legg-2009-tick-tock-bing","legg-2010-goodbye-2010","legg-2011-goodbye-2011","sandberg-bostrom-2011-machine-intelligence-survey","metaculus-5121-capture-2025-01-05","openai-2025-ai-progress-and-recommendations","openai-2025-ai-progress-and-recommendations"],"predicate":["states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction","states_prediction"],"object":[{"kind":"literal","value":"it will be possible, to programme computers, with a storage capacity of about 109, to make them play the imitation game so well that an average interrogator will not have more than 70 per cent chance of making the right identification"},{"kind":"literal","value":"the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted"},{"kind":"literal","value":"a significant advance can be made in one or more of these problems"},{"kind":"literal","value":"within ten years a digital computer will be the world's chess champion, unless the rules bar it from competition"},{"kind":"literal","value":"an ultraintelligent machine will be built and that it will be the last invention that man need make"},{"kind":"literal","value":"computing system exhibiting intelligence at adult human level"},{"kind":"literal","value":"an all-purpose intelligence on a human-scale knowledge base"},{"kind":"literal","value":"machine intelligence exceeding human intelligence"},{"kind":"literal","value":"it will be commonplace to say that the twenty-first century will be dominated by the machine if humanity so chooses—and maybe even if not"},{"kind":"literal","value":"the creation of greater than human intelligence will occur"},{"kind":"literal","value":"this event occurs"},{"kind":"literal","value":"the Singularity"},{"kind":"literal","value":"we will have superhuman artificial intelligence"},{"kind":"literal","value":"the required computational power will be reached"},{"kind":"literal","value":"the required hardware will be available in cheap machines"},{"kind":"literal","value":"robot intelligence will surpass our own"},{"kind":"literal","value":"supercomputers will achieve one human brain capacity by 2010, and personal computers will do so by around 2020"},{"kind":"literal","value":"We achieve one Human Brain capability (2 * 10^16 cps) for $1,000"},{"kind":"literal","value":"the development of human-level machine intelligence"},{"kind":"literal","value":"human-level AI to be reached"},{"kind":"literal","value":"high-level machine intelligence will be developed"},{"kind":"literal","value":"systems will move on to superintelligence"},{"kind":"literal","value":"AI outperforming humans in all tasks"},{"kind":"literal","value":"automating all human jobs"},{"kind":"literal","value":"AI will outperform humans in many activities"},{"kind":"literal","value":"human-level machine intelligence being achieved"},{"kind":"literal","value":"HLMI"},{"kind":"literal","value":"unaided machines outperforming humans in every possible task"},{"kind":"literal","value":"all human occupations becoming fully automatable"},{"kind":"literal","value":"By 2029 no computer - or \"machine intelligence\" - will have passed the Turing Test."},{"kind":"literal","value":"Emergence of the generally agreed upon \"next big thing\" in AI beyond deep learning."},{"kind":"literal","value":"An AI system with an ongoing existence (no day is the repeat of another day as it currently is for all AI systems) at the level of a mouse."},{"kind":"literal","value":"A robot that seems as intelligent, as attentive, and as faithful, as a dog."},{"kind":"literal","value":"transformative AI"},{"kind":"literal","value":"an existential catastrophe of this kind will occur"},{"kind":"literal","value":"transformative AI"},{"kind":"literal","value":"transformative AI"},{"kind":"literal","value":"AGI"},{"kind":"literal","value":"generalist AI systems will outperform human abilities across many critical domains"},{"kind":"literal","value":"AI systems will exceed expert skill level in most domains"},{"kind":"literal","value":"it could arrive"},{"kind":"literal","value":"we will have superintelligence"},{"kind":"literal","value":"powerful AI"},{"kind":"literal","value":"we may see the first AI agents “join the workforce” and materially change the output of companies"},{"kind":"literal","value":"Anyone in 2035 should be able to marshall the intellectual capacity equivalent to everyone in 2025"},{"kind":"literal","value":"powerful AI systems will emerge"},{"kind":"literal","value":"the impact of superhuman AI over the next decade will be enormous, exceeding that of the Industrial Revolution"},{"kind":"literal","value":"powerful AI systems are developed"},{"kind":"literal","value":"we could have AI systems equivalent to a “country of geniuses in a datacenter”"},{"kind":"literal","value":"the arrival of systems that can figure out novel insights"},{"kind":"literal","value":"the arrival of robots that can do tasks in the real world"},{"kind":"literal","value":"powerful AI"},{"kind":"literal","value":"AI is better than humans at essentially everything"},{"kind":"literal","value":"a digital computer will discover and prove an important new mathematical theorem"},{"kind":"literal","value":"a digital computer will write music that will be accepted by critics as possessing considerable aesthetic value"},{"kind":"literal","value":"most theories in psychology will take the form of computer programs, or of qualitative statements about the characteristics of computer programs"},{"kind":"literal","value":"superhuman intelligences are created"},{"kind":"literal","value":"this will change"},{"kind":"literal","value":"roughly human level AGI"},{"kind":"literal","value":"roughly human level AGI"},{"kind":"literal","value":"human level AGI"},{"kind":"literal","value":"HLMI"},{"kind":"literal","value":"When will the first general AI system be devised, tested, and publicly announced?"},{"kind":"literal","value":"AI to be capable of making very small discoveries"},{"kind":"literal","value":"we will have systems that can make more significant discoveries"}],"topic":["ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas","ideas"],"time":[{"kind":"event","date":"1950-10","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"1950-10","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"1955-08-31","note":"Date of the typescript."},{"kind":"event","date":"1958","basis":"publication","note":"Publication date of the text. Operations Research 6 (1), January–February 1958 (the issue names two months). The article is the text of an address of 14 November 1957 (footnote, printed p. 1); the passage dates its own \"present\" to 1957."},{"kind":"event","date":"1965","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"1973","basis":"publication","note":"Publication date of the text. Chapter heading \"(1973)\"; first published in Nature 241 (1973): 507–512. The poll was \"taken last year\", that is in 1972."},{"kind":"event","date":"1973","basis":"publication","note":"Publication date of the text. The report says the predictions were \"made in 1972\"."},{"kind":"event","date":"1973","basis":"publication","note":"Publication date of the text. The report says the predictions were \"made in 1972\"."},{"kind":"event","date":"1989","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"1993-12","basis":"publication","note":"Publication date of the text. NASA conference publication of the VISION-21 symposium (March 1993)."},{"kind":"event","date":"1993-12","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2000-08-31","basis":"publication","note":"Publication date of the text. Revision date of version 1.2.1 (\"Updated: 08/31/2000\"), read in an Internet Archive capture of 2000-12-04; the page was first posted on 1996-11-18 (claimed date; the checked wording is that of version 1.2.1)."},{"kind":"event","date":"1998","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"1998","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"1998","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"1999-12","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2001-01","basis":"publication","note":"Publication date of the text. The live page reads \"January 2001\"; the dataset dates the essay 2001-03-07. An Internet Archive capture of the essay's address of 4 August 2001 (the earliest indexed, of April 2001, redirects to it) prints neither date, read by a reviewer from a different model family on 2026-11-05; no copy read settles the publication date. The dataset's existing records keep 2001-03-07."},{"kind":"event","date":"2001-01","basis":"publication","note":"Publication date of the text. As for the companion record."},{"kind":"event","date":"2011-06-17","basis":"publication","note":"Date of the post. Answers by email, posted by the interviewer (XiXiDu) on LessWrong."},{"kind":"event","date":"2011","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details. Assessment conducted at the AGI-09 conference (2009)."},{"kind":"event","date":"2014","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details. File reads \"forthcoming 2014\"; questionnaire distributed 2012/2013; volume published 2016."},{"kind":"event","date":"2014","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details."},{"kind":"event","date":"2017-05-24","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details. Survey of researchers who published at NIPS and ICML 2015, conducted in 2016."},{"kind":"event","date":"2017-05-24","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details."},{"kind":"event","date":"2017-05-24","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details."},{"kind":"event","date":"2022-06-08","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details. Survey conducted in 2019."},{"kind":"event","date":"2022-08","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details. Live wiki page; wording attested from retrieval on 2026-11-02."},{"kind":"event","date":"2024-01-05","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details. Survey conducted in October 2023."},{"kind":"event","date":"2024-01-05","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details."},{"kind":"event","date":"2002","note":"Year the bet was made. Bet duration \"27 years (02002-02029)\"; the page read is the live bet page."},{"kind":"event","date":"2018-01-01","basis":"publication","note":"Date of the post. Live page; the author publishes yearly reviews of these predictions separately."},{"kind":"event","date":"2018-01-01","basis":"publication","note":"Date of the post."},{"kind":"event","date":"2018-01-01","basis":"publication","note":"Date of the post."},{"kind":"event","date":"2021-09-07","basis":"publication","note":"Date of the post."},{"kind":"event","date":"2022-06-16","basis":"publication","note":"Publication date of the text. Report made public in April 2021 (claimed: the abstract's own statement); the read version carries a May 2022 update."},{"kind":"event","date":"2022-08-02","basis":"publication","note":"Date of the post."},{"kind":"event","date":"2022-08-02","basis":"publication","note":"Date of the post."},{"kind":"event","date":"2024-06","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2023-10-26","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2023-05-22","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2023-07-05","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2024-09-23","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2024-10","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2025-01-05","basis":"publication","note":"Date of the post."},{"kind":"event","date":"2025-02-09","basis":"publication","note":"Date of the post."},{"kind":"event","date":"2025-03-06","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2025-04-03","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2025-04-02","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2025-04","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2025-06-10","basis":"publication","note":"Date of the post."},{"kind":"event","date":"2025-06-10","basis":"publication","note":"Date of the post."},{"kind":"event","date":"2026-01","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"2026-01","basis":"publication","note":"Publication date of the text."},{"kind":"event","date":"1958","basis":"publication","note":"Publication date of the text. Operations Research 6 (1), January–February 1958; the text of an address of 14 November 1957."},{"kind":"event","date":"1958","basis":"publication","note":"Publication date of the text. Operations Research 6 (1), January–February 1958; the text of an address of 14 November 1957."},{"kind":"event","date":"1958","basis":"publication","note":"Publication date of the text. Operations Research 6 (1), January–February 1958; the text of an address of 14 November 1957."},{"kind":"event","date":"1983-01","basis":"publication","note":"Publication date of the text. Omni, January 1983, p. 10 (issue month from the dataset and the hosting site; the page prints no date)."},{"kind":"event","date":"1983-01","basis":"publication","note":"Publication date of the text. Omni, January 1983, p. 10 (issue month from the dataset and the hosting site; the page prints no date)."},{"kind":"event","date":"2009-12-07","basis":"publication","note":"Date of the post. Post date as read by a reviewer from a different model family in a capture of 2009-12-13."},{"kind":"event","date":"2010-12-10","basis":"publication","note":"Date of the post. Post date as reported by a reviewer from a different model family, read in a capture of 2010-12-14."},{"kind":"event","date":"2011-12-31","basis":"publication","note":"Date of the post. Post date as reported by a reviewer from a different model family, read in a capture of 2012-04-06."},{"kind":"event","date":"2011","basis":"publication","note":"Date of the survey report; when the survey was conducted is in the record’s survey details. Poll taken on 16 January 2011 (\"16/1 2011\"); the report is numbered 2011-1."},{"kind":"event","date":"2025-01-05","note":"Date of the copy the figure is read from. The figure is the one displayed on this capture; the question page changes continually."},{"kind":"event","date":"2025-11-06","basis":"publication","note":"Publication date of the text. The capture of 2025-11-07 bears 6 November 2025 (as read by a reviewer from a different model family)."},{"kind":"event","date":"2025-11-06","basis":"publication","note":"Publication date of the text. The capture of 2025-11-07 bears 6 November 2025 (as read by a reviewer from a different model family)."}],"attributes":[{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Alan Turing","entityIds":["alan-turing"]},"event":{"type":"quote","text":"it will be possible, to programme computers, with a storage capacity of about 109, to make them play the imitation game so well that an average interrogator will not have more than 70 per cent chance of making the right identification","note":"The sentence continues \"after five minutes of questioning\" (second quotation)."},"horizon":{"asStated":"in about fifty years' time","words":"in about fifty years' time","kind":"relative","interval":{"amount":"about fifty","unit":"years","from":"the source date"},"precision":"approximate"},"hedge":"Introduced as the author's belief (\"I believe that\")."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Alan Turing","entityIds":["alan-turing"]},"event":{"type":"quote","text":"the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted"},"horizon":{"asStated":"at the end of the century","words":"at the end of the century","kind":"period","precision":"end of a century"},"hedge":"Introduced as the author's belief (\"I believe that\")."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"persons","label":"John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon (the proposal's signatories)","entityIds":["john-mccarthy","marvin-minsky","nathaniel-rochester","claude-e-shannon"]},"event":{"type":"quote","text":"a significant advance can be made in one or more of these problems"},"horizon":{"asStated":"for a summer (a \"2 month, 10 man study\" during the summer of 1956)","words":"work on it together for a summer","kind":"relative","interval":{"amount":"one","unit":"summer","from":"the proposed study"},"precision":"season"},"conditions":"if a carefully selected group of scientists work on it together for a summer","hedge":"Introduced as \"We think that\"."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"persons","label":"Herbert A. Simon and Allen Newell"},"event":{"type":"quote","text":"within ten years a digital computer will be the world's chess champion, unless the rules bar it from competition"},"horizon":{"asStated":"within ten years","words":"within ten years","kind":"relative","interval":{"amount":"ten","unit":"years","from":"occasion"},"precision":"year"},"occasion":{"date":"1957-11-14","kind":"address","note":"Banquet address at the Twelfth National Meeting of the Operations Research Society of America, Pittsburgh (footnote, printed p. 1); the passage says \"As of the present—1957\"."},"confidence":{"asStated":"\"I am willing to make the following predictions\"","words":"I am willing to make the following predictions"},"conditions":"unless the rules bar it from competition"}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"I. J. Good","entityIds":["ij-good"]},"event":{"type":"quote","text":"an ultraintelligent machine will be built and that it will be the last invention that man need make"},"horizon":{"asStated":"within the twentieth century","words":"within the twentieth century","kind":"period","precision":"century"},"probability":{"asStated":"\"more probable than not\"","words":"It is more probable than not"}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Opinion poll reported by Donald Michie, \"Machines and the theory of intelligence\" (Nature, 1973)"},"event":{"type":"quote","text":"computing system exhibiting intelligence at adult human level"},"horizon":{"asStated":"Given only in Fig. 11.4, as the frequency of answers over \"Estimated number of years before\" the event, in bands from 10 to more than 50 years","kind":"figure","precision":"figure; values not read off"},"occasion":{"date":"1972","kind":"poll","note":"The text says the poll was \"taken last year\"; the chapter is dated 1973."},"survey":{"respondents":"sixty-seven British and American computer scientists","conducted":"1972 (\"last year\")","question":"Estimated number of years before the event (figure caption)"},"attested":{"date":"1986","note":"Wording read in the 1986 second edition of On Machine Intelligence (chapter 11)."}}},{"prediction":{"form":"reported-statement","predictor":{"kind":"unnamed","label":"Unnamed correspondents (\"able and respected scientists\")","reporter":{"label":"James Lighthill"}},"event":{"type":"quote","text":"an all-purpose intelligence on a human-scale knowledge base"},"horizon":{"asStated":"in the nineteen-eighties","words":"in the nineteen-eighties","kind":"period","precision":"decade"},"hedge":"Worded as \"possibilities\"."}},{"prediction":{"form":"reported-statement","predictor":{"kind":"unnamed","label":"Unnamed correspondents (\"able and respected scientists\")","reporter":{"label":"James Lighthill"}},"event":{"type":"quote","text":"machine intelligence exceeding human intelligence"},"horizon":{"asStated":"by the year 2000","words":"by the year 2000","kind":"absolute","years":[2000],"precision":"year"},"hedge":"Worded as \"awe-inspiring possibilities suggest themselves\"."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Hugo de Garis"},"event":{"type":"quote","text":"it will be commonplace to say that the twenty-first century will be dominated by the machine if humanity so chooses—and maybe even if not"},"horizon":{"asStated":"In 20 years time","words":"In 20 years time","kind":"relative","interval":{"amount":"20","unit":"years","from":"the source date"},"precision":"year"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Vernor Vinge","entityIds":["vernor-vinge"]},"event":{"type":"quote","text":"the creation of greater than human intelligence will occur"},"horizon":{"asStated":"during the next thirty years","words":"during the next thirty years","kind":"relative","interval":{"amount":"thirty","unit":"years","from":"the source date"},"precision":"year"},"hedge":"Introduced as \"I believe that\"."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Vernor Vinge","entityIds":["vernor-vinge"]},"event":{"type":"quote","text":"this event occurs","note":"\"This event\" is the creation of greater than human intelligence (preceding sentence)."},"horizon":{"asStated":"I'll be surprised if this event occurs before 2005 or after 2030","words":"before 2005 or after 2030","kind":"absolute","years":[2005,2030],"relation":"window","precision":"year"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Eliezer Yudkowsky","entityIds":["eliezer-yudkowsky"]},"event":{"type":"paraphrase","text":"the Singularity","note":"Paraphrase: the quoted sentence says \"it seems like a reasonable upper bound\"; \"it\" refers back to the subject of the section \"Sooner Than You Think\", the Singularity. The horizon words lie in the quotation."},"horizon":{"asStated":"2035 \"seems like a reasonable upper bound\"; \"The lower bound? Thirty seconds.\"","words":"The most we can say about 2035 is that it seems like a reasonable upper bound","kind":"absolute","years":[2035],"relation":"upper bound","precision":"year"},"conditions":"\"given the current rate of progress\"","hedge":"\"The most we can say\"."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Nick Bostrom","entityIds":["nick-bostrom"]},"event":{"type":"quote","text":"we will have superhuman artificial intelligence"},"horizon":{"asStated":"within the first third of the next century","words":"within the first third of the next century","kind":"period","precision":"third of a century"},"hedge":"Worded as \"the case for believing that\".","attested":{"date":"2026-11-02","note":"Wording read on the author-hosted page on 2026-11-02 and again on 2026-11-03; a reviewer from a different model family found, on 2026-11-05, the earliest located copy, of December 2000 but did not transcribe this passage from it. The source date is the document's own; the wording is attested only from the retrieval date."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Nick Bostrom","entityIds":["nick-bostrom"]},"event":{"type":"quote","text":"the required computational power will be reached"},"horizon":{"asStated":"sometime between 2004 and 2008, depending on whether we assume a doubling time of 12 or 18 months","words":"sometime between 2004 and 2008","kind":"absolute","years":[2004,2008],"relation":"window","precision":"year"},"conditions":"If we need 100 Tops to simulate the human brain; doubling time of 12 or 18 months.","attested":{"date":"2000-12-10","note":"The quoted sentence is in the earliest located Internet Archive capture of the page (20001210012600), read by a reviewer from a different model family on 2026-11-05. That copy is titled \"(1997)\", \"[Revised 25 October, 1998, and provided with a postscript]\" and \"[A second postscript added 28 August, 2000]\": it attests the wording in December 2000, not in 1998."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Hans Moravec","entityIds":["hans-moravec"]},"event":{"type":"quote","text":"the required hardware will be available in cheap machines","note":"Hardware \"to match general intellectual performance of the human brain\" (abstract)."},"horizon":{"asStated":"in the 2020s","words":"in the 2020s","kind":"period","precision":"decade"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Hans Moravec","entityIds":["hans-moravec"]},"event":{"type":"quote","text":"robot intelligence will surpass our own"},"horizon":{"asStated":"well before 2050","words":"well before 2050","kind":"absolute","years":[2050],"relation":"before","precision":"year"},"hedge":"\"It suggests that\"."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Ray Kurzweil"},"event":{"type":"quote","text":"supercomputers will achieve one human brain capacity by 2010, and personal computers will do so by around 2020"},"horizon":{"asStated":"by 2010 (supercomputers); by around 2020 (personal computers)","words":"by around 2020","kind":"absolute","years":[2010,2020],"precision":"year"},"points":[{"horizon":"by 2010","subject":"supercomputers"},{"horizon":"by around 2020","subject":"personal computers"}],"attested":{"date":"2001-08-04","note":"The quoted sentence is in an Internet Archive capture of the essay's address on kurzweilai.net of 4 August 2001 (20010804224404), read by a reviewer from a different model family on 2026-11-05. That copy is headed \"The Singularity is Near\" and \"A Book Précis by Raymond Kurzweil\", with \"Copyright ( C ) Raymond Kurzweil 2001\"; it attests the sentence at that address in August 2001, not the title or version read on the live page. The earliest indexed capture (April 2001) redirects to it."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Ray Kurzweil"},"event":{"type":"quote","text":"We achieve one Human Brain capability (2 * 10^16 cps) for $1,000"},"horizon":{"asStated":"around the year 2023","words":"around the year 2023","kind":"absolute","years":[2023],"relation":"around","precision":"year"},"attested":{"date":"2026-11-02","note":"Wording read on the page on the author's site on 2026-11-02 and again on 2026-11-03; a reviewer from a different model family found, on 2026-11-05, the copy of August 2001 but did not transcribe this passage from it. The source date is the document's own; the wording is attested only from the retrieval date."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Shane Legg","entityIds":["shane-legg"]},"event":{"type":"quote","text":"the development of human-level machine intelligence"},"horizon":{"asStated":"10%: 2018; 50%: 2028; 90%: 2050","words":"2018, 2028, 2050","kind":"quantile","years":[2018,2028,2050],"precision":"year"},"probability":{"asStated":"10%, 50% and 90% (the question's levels)","words":"10%/50%/90% chance"},"points":[{"probability":"10%","horizon":"2018"},{"probability":"50%","horizon":"2028"},{"probability":"90%","horizon":"2050"}],"conditions":"Assuming no global catastrophe halts progress (the question's wording)."}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Baum, Goertzel and Goertzel, expert assessment at AGI-09"},"event":{"type":"quote","text":"human-level AI to be reached"},"horizon":{"asStated":"within upcoming decades","words":"within upcoming decades","kind":"vague","precision":"decades"},"survey":{"respondents":"participants of the AGI-09 conference","conducted":"2009"}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Müller and Bostrom, survey of expert opinion"},"event":{"type":"quote","text":"high-level machine intelligence will be developed"},"horizon":{"asStated":"50%: around 2040-2050; 90%: by 2075 (median estimates)","words":"around 2040-2050, rising to a nine in ten chance by 2075","kind":"quantile","years":[2040,2050,2075],"precision":"year"},"probability":{"asStated":"\"a one in two chance\"; \"a nine in ten chance\"","words":"a one in two chance"},"points":[{"probability":"a one in two chance","horizon":"around 2040-2050"},{"probability":"a nine in ten chance","horizon":"by 2075"}],"conditions":"The question assumes \"that human scientific activity continues without major negative disruption\".","survey":{"respondents":"four groups of experts","conducted":"2012/2013","statistic":"median of respondents' estimates"}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Müller and Bostrom, survey of expert opinion"},"event":{"type":"quote","text":"systems will move on to superintelligence"},"horizon":{"asStated":"in less than 30 years thereafter (after high-level machine intelligence); the question asked for probabilities \"within (2 years / 30 years) thereafter\"","words":"in less than 30 years thereafter","kind":"relative","interval":{"amount":"less than 30","unit":"years","from":"high-level machine intelligence"},"precision":"year"},"probability":{"asStated":"Not in the abstract. The body reports, for all respondents, a median of 75% within 30 years and 10% within 2 years (means 62% and 19%), in a table (PDF p. 12), and summarises it as \"a significant probability\"","words":"a significant probability for superintelligence within 30 years"},"conditions":"The question assumes that high-level machine intelligence \"will at some point exist\".","survey":{"respondents":"four groups of experts","conducted":"2012/2013"}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Grace et al., survey of machine learning researchers (2016)"},"event":{"type":"quote","text":"AI outperforming humans in all tasks"},"horizon":{"asStated":"50% chance in 45 years","words":"in 45 years","kind":"quantile","interval":{"amount":"45","unit":"years","from":"the survey (2016)"},"precision":"year"},"probability":{"asStated":"50%","words":"50% chance"},"survey":{"respondents":"researchers who published at NIPS and ICML 2015","conducted":"2016","statistic":"aggregate forecast (mean of fitted individual distributions)"}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Grace et al., survey of machine learning researchers (2016)"},"event":{"type":"quote","text":"automating all human jobs"},"horizon":{"asStated":"50% chance in 120 years","words":"in 120 years","kind":"quantile","interval":{"amount":"120","unit":"years","from":"the survey (2016)"},"precision":"year"},"probability":{"asStated":"50%","words":"50% chance"},"survey":{"respondents":"researchers who published at NIPS and ICML 2015","conducted":"2016"}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Grace et al., survey of machine learning researchers (2016)"},"event":{"type":"quote","text":"AI will outperform humans in many activities"},"horizon":{"asStated":"in the next ten years (examples: translating languages by 2024, writing high-school essays by 2026, driving a truck by 2027)","words":"in the next ten years","kind":"relative","interval":{"amount":"ten","unit":"years","from":"the survey"},"precision":"year"},"survey":{"respondents":"researchers who published at NIPS and ICML 2015","conducted":"2016"}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Zhang et al., survey of AI/ML researchers (2019)"},"event":{"type":"quote","text":"human-level machine intelligence being achieved","note":"For the cross-sectional sample that gives 2060, the report defines human-level machine intelligence as machines collectively performing \"almost all tasks (>90% of all tasks) that are economically relevant\" better than the median paid human (second quotation)."},"horizon":{"asStated":"50% likelihood by 2060","words":"by 2060","kind":"quantile","years":[2060],"precision":"year"},"probability":{"asStated":"50% likelihood","words":"50% likelihood"},"survey":{"respondents":"AI/ML researchers","conducted":"2019","statistic":"aggregate"}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"AI Impacts, 2022 Expert Survey on Progress in AI"},"event":{"type":"quote","text":"HLMI","note":"High-level machine intelligence as the survey defines it."},"horizon":{"asStated":"aggregate forecast time to a 50% chance: 37 years, i.e. 2059","words":"37 years, i.e. 2059","kind":"quantile","years":[2059],"precision":"year"},"probability":{"asStated":"50% chance","words":"50% chance"},"conditions":"Conditional on \"human scientific activity continu[ing] without major negative disruption\" (the page's wording of the question's condition).","survey":{"conducted":"2022","statistic":"aggregate forecast"},"attested":{"date":"2022-08-04","note":"Internet Archive capture 20220804162728 of the results post at aiimpacts.org, read by a reviewer from a different model family: it carries the \"37 years, i.e. 2059\" summary and the 36.6-year figure with its condition. The record's quotations are from the live wiki page."}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Grace et al., survey of AI authors (2023)"},"event":{"type":"quote","text":"unaided machines outperforming humans in every possible task"},"horizon":{"asStated":"10% by 2027, and 50% by 2047","words":"10% by 2027, and 50% by 2047","kind":"quantile","years":[2027,2047],"precision":"year"},"probability":{"asStated":"10%; 50%","words":"10% by 2027"},"points":[{"probability":"10%","horizon":"by 2027"},{"probability":"50%","horizon":"by 2047"}],"conditions":"If science continues undisrupted.","survey":{"conducted":"2023","statistic":"aggregate forecast"}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Grace et al., survey of AI authors (2023)"},"event":{"type":"quote","text":"all human occupations becoming fully automatable"},"horizon":{"asStated":"10% by 2037, and 50% as late as 2116","words":"10% by 2037, and 50% as late as 2116","kind":"quantile","years":[2037,2116],"precision":"year"},"probability":{"asStated":"10%; 50%","words":"10% by 2037"},"points":[{"probability":"10%","horizon":"by 2037"},{"probability":"50%","horizon":"as late as 2116"}],"survey":{"conducted":"2023","statistic":"aggregate forecast"}}},{"prediction":{"form":"public-bet","predictor":{"kind":"persons","label":"Mitchell Kapor (predictor) and Ray Kurzweil (challenger), on Long Bets"},"event":{"type":"quote","text":"By 2029 no computer - or \"machine intelligence\" - will have passed the Turing Test."},"horizon":{"asStated":"By 2029","words":"By 2029","kind":"absolute","years":[2029],"relation":"by","precision":"year"},"bet":{"platform":"Long Bets (The Long Now Foundation)","parties":[{"label":"Mitchell Kapor","position":"predictor: that by 2029 no computer will have passed the Turing Test"},{"label":"Ray Kurzweil","position":"challenger: the contrary"}],"stakes":"$20,000, to the Electronic Frontier Foundation if Kapor wins or the Kurzweil Foundation if Kurzweil wins (as stated on the page)"},"attested":{"date":"2026-11-02","note":"Wording read on the live bet page on 2026-11-02 and again on 2026-11-03. The capture of 7 June 2002 words the bet differently (see the qualification), so it does not attest this wording. The source date is the bet's own; the wording is attested only from the retrieval date."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Rodney Brooks"},"event":{"type":"quote","text":"Emergence of the generally agreed upon \"next big thing\" in AI beyond deep learning."},"horizon":{"asStated":"NET 2023, BY 2027 (\"No Earlier Than\" 2023, \"By\" 2027)","words":"NET 2023 BY 2027","kind":"absolute","years":[2023,2027],"relation":"window","precision":"year"},"attested":{"date":"2018-01-02","note":"The quoted wording, and the \"NET\" and \"BY\" conventions, are in the earliest located Internet Archive capture of January–February 2018 (20180102073550), dated \"January 1, 2018\", read by a reviewer from a different model family on 2026-11-05."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Rodney Brooks"},"event":{"type":"quote","text":"An AI system with an ongoing existence (no day is the repeat of another day as it currently is for all AI systems) at the level of a mouse."},"horizon":{"asStated":"NET 2030 (\"No Earlier Than\")","words":"NET 2030","kind":"absolute","years":[2030],"relation":"no earlier than","precision":"year"},"attested":{"date":"2018-01-02","note":"The quoted wording, and the \"NET\" and \"BY\" conventions, are in the earliest located Internet Archive capture of January–February 2018 (20180102073550), dated \"January 1, 2018\", read by a reviewer from a different model family on 2026-11-05."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Rodney Brooks"},"event":{"type":"quote","text":"A robot that seems as intelligent, as attentive, and as faithful, as a dog."},"horizon":{"asStated":"NET 2048 (\"No Earlier Than\")","words":"NET 2048","kind":"absolute","years":[2048],"relation":"no earlier than","precision":"year"},"attested":{"date":"2018-01-02","note":"The quoted wording, and the \"NET\" and \"BY\" conventions, are in the earliest located Internet Archive capture of January–February 2018 (20180102073550), dated \"January 1, 2018\", read by a reviewer from a different model family on 2026-11-05."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Holden Karnofsky","entityIds":["holden-karnofsky"]},"event":{"type":"quote","text":"transformative AI","note":"Defined in the next sentence as \"AI powerful enough to bring us into a new, qualitatively different future\"."},"horizon":{"asStated":"more than 10% within 15 years (by 2036); ~50% within 40 years (by 2060); ~2/3 this century (by 2100)","words":"within 15 years (by 2036)","kind":"quantile","years":[2036,2060,2100],"precision":"year"},"probability":{"asStated":"more than a 10% chance; a ~50% chance; a ~2/3 chance","words":"more than a 10% chance"},"points":[{"probability":"more than a 10% chance","horizon":"within 15 years (by 2036)"},{"probability":"a ~50% chance","horizon":"within 40 years (by 2060)"},{"probability":"a ~2/3 chance","horizon":"this century (by 2100)"}]}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Joseph Carlsmith"},"event":{"type":"quote","text":"an existential catastrophe of this kind will occur","note":"\"Of this kind\": from power-seeking AI systems, as the report's six-premise argument specifies."},"horizon":{"asStated":"by 2070","words":"by 2070","kind":"absolute","years":[2070],"relation":"by","precision":"year"},"probability":{"asStated":"~5% (April 2021); \">10%\" (May 2022 update)","words":"~5%"},"points":[{"probability":"~5%","horizon":"by 2070","version":"the report as made public in April 2021","date":"2021-04"},{"probability":">10%","horizon":"by 2070","version":"May 2022 update (the horizon is carried over from \"my estimate here\")","date":"2022-05"}]}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Ajeya Cotra","entityIds":["ajeya-cotra"]},"event":{"type":"quote","text":"transformative AI"},"horizon":{"asStated":"~15% by 2036; a median of ~2050","words":"A median of ~2050","kind":"quantile","years":[2036,2050],"precision":"year"},"probability":{"asStated":"~15%; median","words":"Roughly ~15% probability"},"points":[{"probability":"~15% probability","horizon":"by 2036"},{"probability":"A median","horizon":"~2050"}],"hedge":"\"Roughly\"; the next sentence says these were \"roughly close to my all-things-considered probabilities at the time\".","retrospective":{"estimateOf":"2020-09","note":"The 2022 post reports the bottom line of the draft report posted in September 2020. The 2020 report (a linked document) was not read for this record."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Ajeya Cotra","entityIds":["ajeya-cotra"]},"event":{"type":"quote","text":"transformative AI"},"horizon":{"asStated":"~15% by 2030; ~35% by 2036; a median of ~2040; ~60% by 2050","words":"A median of ~2040","kind":"quantile","years":[2030,2036,2040,2050],"precision":"year"},"probability":{"asStated":"~15%; ~35%; median; ~60%","words":"~15% probability by 2030"},"points":[{"probability":"~15% probability","horizon":"by 2030"},{"probability":"~35% probability","horizon":"by 2036"},{"probability":"A median","horizon":"~2040"},{"probability":"~60% probability","horizon":"by 2050"}],"hedge":"\"something roughly like this\"."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Leopold Aschenbrenner","entityIds":["leopold-aschenbrenner"]},"event":{"type":"quote","text":"AGI"},"horizon":{"asStated":"by 2027","words":"AGI by 2027","kind":"absolute","years":[2027],"relation":"by","precision":"year"},"confidence":{"asStated":"\"strikingly plausible\"","words":"strikingly plausible"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"persons","label":"Yoshua Bengio, Geoffrey Hinton, Andrew Yao and co-authors","entityIds":["yoshua-bengio"]},"event":{"type":"quote","text":"generalist AI systems will outperform human abilities across many critical domains"},"horizon":{"asStated":"within this decade or the next","words":"within this decade or the next","kind":"period","precision":"decade"},"confidence":{"asStated":"\"we must take seriously the possibility\"","words":"we must take seriously the possibility"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"persons","label":"Sam Altman, Greg Brockman and Ilya Sutskever (OpenAI)","entityIds":["sam-altman","greg-brockman","ilya-sutskever"]},"event":{"type":"quote","text":"AI systems will exceed expert skill level in most domains"},"horizon":{"asStated":"within the next ten years","words":"within the next ten years","kind":"relative","interval":{"amount":"ten","unit":"years","from":"the source date"},"precision":"year"},"confidence":{"asStated":"\"it's conceivable\"","words":"it’s conceivable"},"attested":{"date":"2023-05-22","note":"Internet Archive capture 20230522175341 of the original address, read by a reviewer from a different model family."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"organisation","label":"OpenAI (post by Jan Leike and Ilya Sutskever)","entityIds":["openai"]},"event":{"type":"quote","text":"it could arrive","note":"\"It\" is superintelligence."},"horizon":{"asStated":"this decade","words":"this decade","kind":"period","precision":"decade"},"confidence":{"asStated":"\"we believe it could\"","words":"we believe it could arrive"},"attested":{"date":"2023-07-05","note":"Internet Archive capture 20230705170451 of the original address, read by a reviewer from a different model family."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Sam Altman","entityIds":["sam-altman"]},"event":{"type":"quote","text":"we will have superintelligence"},"horizon":{"asStated":"in a few thousand days (!); it may take longer","words":"in a few thousand days","kind":"relative","interval":{"amount":"a few thousand","unit":"days","from":"the source date"},"precision":"vague"},"confidence":{"asStated":"\"It is possible\"; \"I'm confident we'll get there\"","words":"It is possible"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Dario Amodei","entityIds":["dario-amodei"]},"event":{"type":"quote","text":"powerful AI","note":"Defined in the essay (\"powerful AI\")."},"horizon":{"asStated":"as early as 2026, though there are also ways it could take much longer","words":"as early as 2026","kind":"absolute","years":[2026],"relation":"as early as","precision":"year"},"confidence":{"asStated":"\"I think it could\"","words":"I think it could come"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Sam Altman","entityIds":["sam-altman"]},"event":{"type":"quote","text":"we may see the first AI agents “join the workforce” and materially change the output of companies"},"horizon":{"asStated":"in 2025","words":"in 2025","kind":"absolute","years":[2025],"relation":"in","precision":"year"},"confidence":{"asStated":"\"We believe that ... we may\"","words":"we may see"}}},{"prediction":{"form":"stated-forecast","mode":"forecast-or-aim","predictor":{"kind":"person","label":"Sam Altman","entityIds":["sam-altman"]},"event":{"type":"quote","text":"Anyone in 2035 should be able to marshall the intellectual capacity equivalent to everyone in 2025"},"horizon":{"asStated":"in 2035","words":"in 2035","kind":"absolute","years":[2035],"relation":"in","precision":"year"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"organisation","label":"Anthropic","entityIds":["anthropic"]},"event":{"type":"quote","text":"powerful AI systems will emerge"},"horizon":{"asStated":"in late 2026 or early 2027","words":"in late 2026 or early 2027","kind":"period","precision":"part of a year"},"confidence":{"asStated":"\"we expect\"","words":"we expect"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"persons","label":"Daniel Kokotajlo, Scott Alexander, Thomas Larsen, Eli Lifland and Romeo Dean (AI 2027)","entityIds":["daniel-kokotajlo"]},"event":{"type":"quote","text":"the impact of superhuman AI over the next decade will be enormous, exceeding that of the Industrial Revolution"},"horizon":{"asStated":"over the next decade","words":"over the next decade","kind":"relative","interval":{"amount":"a","unit":"decade","from":"the source date"},"precision":"decade"},"attested":{"date":"2025-04-03","note":"The quoted sentences are in the earliest located Internet Archive capture of April 2025 (20250403163539), dated \"April 3rd 2025\", read by a reviewer from a different model family on 2026-11-05."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"organisation","label":"Google DeepMind (paper by Rohin Shah and co-authors)","entityIds":["google-deepmind"]},"event":{"type":"quote","text":"powerful AI systems are developed"},"horizon":{"asStated":"by 2030","words":"by 2030","kind":"absolute","years":[2030],"relation":"by","precision":"year"},"confidence":{"asStated":"\"we find it plausible\"","words":"we find it plausible"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Dario Amodei","entityIds":["dario-amodei"]},"event":{"type":"quote","text":"we could have AI systems equivalent to a “country of geniuses in a datacenter”"},"horizon":{"asStated":"as soon as 2026 or 2027","words":"as soon as 2026 or 2027","kind":"absolute","years":[2026,2027],"relation":"as soon as","precision":"year"},"confidence":{"asStated":"\"could\"","words":"we could have"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Sam Altman","entityIds":["sam-altman"]},"event":{"type":"quote","text":"the arrival of systems that can figure out novel insights"},"horizon":{"asStated":"2026","words":"2026 will likely see","kind":"absolute","years":[2026],"relation":"in","precision":"year"},"confidence":{"asStated":"\"will likely\"","words":"will likely see"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Sam Altman","entityIds":["sam-altman"]},"event":{"type":"quote","text":"the arrival of robots that can do tasks in the real world"},"horizon":{"asStated":"2027","words":"2027 may see","kind":"absolute","years":[2027],"relation":"in","precision":"year"},"confidence":{"asStated":"\"may\"","words":"2027 may see"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Dario Amodei","entityIds":["dario-amodei"]},"event":{"type":"quote","text":"powerful AI"},"horizon":{"asStated":"as little as 1–2 years away, although it could also be considerably further out","words":"as little as 1–2 years away","kind":"relative","interval":{"amount":"1–2","unit":"years","from":"the source date"},"relation":"as little as","precision":"year"},"confidence":{"asStated":"\"could be\"","words":"powerful AI could be"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Dario Amodei","entityIds":["dario-amodei"]},"event":{"type":"quote","text":"AI is better than humans at essentially everything"},"horizon":{"asStated":"it cannot possibly be more than a few years before","words":"it cannot possibly be more than a few years before","kind":"vague","precision":"years"},"conditions":"If the exponential continues (\"which is not certain, but now has a decade-long track record supporting it\")."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"persons","label":"Herbert A. Simon and Allen Newell"},"event":{"type":"quote","text":"a digital computer will discover and prove an important new mathematical theorem"},"horizon":{"asStated":"within ten years","words":"within ten years","kind":"relative","interval":{"amount":"ten","unit":"years","from":"occasion"},"precision":"year"},"occasion":{"date":"1957-11-14","kind":"address","note":"Banquet address at the Twelfth National Meeting of the Operations Research Society of America, Pittsburgh (footnote, printed p. 1); the passage says \"As of the present—1957\"."},"confidence":{"asStated":"\"I am willing to make the following predictions\"","words":"I am willing to make the following predictions"},"conditions":"Grounded in \"On the basis of these developments, and the speed with which research in this field is progressing\" (the stated basis of all four predictions, not a condition)."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"persons","label":"Herbert A. Simon and Allen Newell"},"event":{"type":"quote","text":"a digital computer will write music that will be accepted by critics as possessing considerable aesthetic value"},"horizon":{"asStated":"within ten years","words":"within ten years","kind":"relative","interval":{"amount":"ten","unit":"years","from":"occasion"},"precision":"year"},"occasion":{"date":"1957-11-14","kind":"address","note":"Banquet address at the Twelfth National Meeting of the Operations Research Society of America, Pittsburgh (footnote, printed p. 1); the passage says \"As of the present—1957\"."},"confidence":{"asStated":"\"I am willing to make the following predictions\"","words":"I am willing to make the following predictions"},"conditions":"Grounded in \"On the basis of these developments, and the speed with which research in this field is progressing\" (the stated basis of all four predictions, not a condition)."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"persons","label":"Herbert A. Simon and Allen Newell"},"event":{"type":"quote","text":"most theories in psychology will take the form of computer programs, or of qualitative statements about the characteristics of computer programs"},"horizon":{"asStated":"within ten years","words":"within ten years","kind":"relative","interval":{"amount":"ten","unit":"years","from":"occasion"},"precision":"year"},"occasion":{"date":"1957-11-14","kind":"address","note":"Banquet address at the Twelfth National Meeting of the Operations Research Society of America, Pittsburgh (footnote, printed p. 1); the passage says \"As of the present—1957\"."},"confidence":{"asStated":"\"I am willing to make the following predictions\"","words":"I am willing to make the following predictions"},"conditions":"Grounded in \"On the basis of these developments, and the speed with which research in this field is progressing\" (the stated basis of all four predictions, not a condition)."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Vernor Vinge","entityIds":["vernor-vinge"]},"event":{"type":"quote","text":"superhuman intelligences are created","note":"The column says the future it describes \"will happen as soon as superhuman intelligences are created\"; \"that should be\" refers back to that creation."},"horizon":{"asStated":"between 20 and 100 years from now","words":"between 20 and 100 years from now","kind":"relative","interval":{"amount":"between 20 and 100","unit":"years","from":"the source date"},"relation":"window","precision":"decades"},"confidence":{"asStated":"\"should be\"","words":"should be"},"conditions":"Grounded in \"progress in computer and biological sciences\" (the column's words; the word before \"progress\" is not quoted, see the copy note)."}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Vernor Vinge","entityIds":["vernor-vinge"]},"event":{"type":"quote","text":"this will change","note":"\"This\" is the state described in the previous sentence: \"little progress in automating the distinct prerogatives of the human mind\". The column goes on to programs for symbolic processing and students' hand computers."},"horizon":{"asStated":"During the next 20 years","words":"During the next 20 years","kind":"relative","interval":{"amount":"20","unit":"years","from":"the source date"},"relation":"within","precision":"year"}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Shane Legg","entityIds":["shane-legg"]},"event":{"type":"paraphrase","text":"roughly human level AGI","note":"The distribution's subject, named in the paragraph's first sentence: \"my prediction for the last 10 years has been for roughly human level AGI in the year 2025\"; the post goes on: \"This year I've tried to come up with something a bit more precise.\""},"horizon":{"asStated":"mode \"about 2025\"; expected value \"a bit higher at 2028\"; 90% credibility region \"2018 to 2036\"","words":"my mode is about 2025, my expected value is actually a bit higher at 2028","kind":"quantile","years":[2025,2028,2018,2036],"precision":"year"},"probability":{"asStated":"mode; expected value; \"90% credibility region\"","words":"90% credibility region"},"points":[{"probability":"mode","horizon":"about 2025"},{"probability":"expected value","horizon":"a bit higher at 2028"},{"probability":"90% credibility region","horizon":"2018 to 2036"}],"attested":{"date":"2009-12-13","note":"Internet Archive capture 20091213040508, read by a reviewer from a different model family."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Shane Legg","entityIds":["shane-legg"]},"event":{"type":"paraphrase","text":"roughly human level AGI","note":"The paragraph's subject: \"My longest running prediction, since 1999, has been the time until roughly human level AGI.\""},"horizon":{"asStated":"log-normal distribution with a mean of 2028, and a mode of 2025","words":"a mean of 2028, and a mode of 2025","kind":"quantile","years":[2028,2025],"precision":"year"},"probability":{"asStated":"mean; mode","words":"log-normal distribution"},"points":[{"probability":"a mean","horizon":"2028"},{"probability":"a mode","horizon":"2025"}],"attested":{"date":"2010-12-14","note":"Internet Archive capture 20101214110948, read by a reviewer from a different model family."},"retrospective":{"estimateOf":"2009-12","note":"The post reports the distribution given \"last year\" (the post of 7 December 2009) and keeps it: \"keep the prediction unmoved at 2028\". The record dates the restatement by this post; the 2009 estimate is recorded separately."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"person","label":"Shane Legg","entityIds":["shane-legg"]},"event":{"type":"paraphrase","text":"human level AGI","note":"The sentence before the quotation: \"I’ve decided to once again leave my prediction for when human level AGI will arrive unchanged.\" The quotation follows \"That is,\"."},"horizon":{"asStated":"log-normal distribution with a mean of 2028 and a mode of 2025","words":"a mean of 2028 and a mode of 2025","kind":"quantile","years":[2028,2025],"precision":"year"},"probability":{"asStated":"mean; mode (of a log-normal distribution)","words":"a mean of 2028 and a mode of 2025"},"points":[{"probability":"a mean","horizon":"2028"},{"probability":"a mode","horizon":"2025"}],"conditions":"On the assumption that nothing catastrophic, such as a nuclear war, happens (paraphrased).","attested":{"date":"2012-04-06","note":"Internet Archive capture 20120406082032, read by a reviewer from a different model family."}}},{"prediction":{"form":"expert-survey","predictor":{"kind":"survey","label":"Sandberg and Bostrom, informal poll at the FHI Winter Intelligence conference (FHI Technical Report 2011-1)"},"event":{"type":"quote","text":"HLMI","note":"Human-level machine intelligence, which the appended questionnaire defines as one that can substitute for humans in virtually all cognitive tasks, including those requiring scientific creativity, common sense or social skills (paraphrased)."},"horizon":{"asStated":"median estimates: 10% chance of HLMI 2028; 50% chance 2050; 90% chance 2150","words":"was 2050","kind":"quantile","years":[2028,2050,2150],"precision":"year"},"probability":{"asStated":"10%; 50%; 90%","words":"50% chance of HLMI"},"points":[{"probability":"10% chance of HLMI","horizon":"was 2028"},{"probability":"50% chance of HLMI","horizon":"was 2050"},{"probability":"90% chance of HLMI","horizon":"was 2150"}],"conditions":"Assuming no global catastrophe halts progress (the question as put to respondents).","survey":{"respondents":"participants of the FHI Winter Intelligence conference on machine intelligence; 35 surveys were collected, a response rate of about 41% according to the report","conducted":"2011-01-16","statistic":"median estimate","question":"By what year respondents would assign a 10%, 50% or 90% chance of the development of human-level machine intelligence, assuming no global catastrophe halts progress (paraphrased from question 1)."},"attested":{"date":"2016-12-24","note":"Internet Archive capture 20161224100323 of the report PDF, read by a reviewer from a different model family."}}},{"prediction":{"form":"platform-question","predictor":{"kind":"platform","label":"Metaculus community prediction, question 5121"},"event":{"type":"quote","text":"When will the first general AI system be devised, tested, and publicly announced?","note":"The question's title on this capture; its resolution criteria (about 3,000 characters on the capture) define \"general AI system\" and changed over time. The record points to the capture for them."},"horizon":{"asStated":"2031-06-14 (the community prediction as displayed; the community table's \"median\", between \"lower 25%\" 2027-05-10 and \"upper 75%\" 2041-08-31)","words":"2031-06-14","kind":"absolute","years":[2031],"precision":"day"},"probability":{"asStated":"median (community table)","words":"median"},"capture":{"date":"2025-01-05","timestamp":"20250105212836","url":"https://web.archive.org/web/20250105212836/https://www.metaculus.com/questions/5121/date-of-artificial-general-intelligence/","note":"Internet Archive capture, read by a reviewer from a different model family. Of 16 dated captures of questions 5121 and 3479 that this reviewer read, this is the only one with a readable community figure in its first-of-January slot."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"organisation","label":"OpenAI","entityIds":["openai"]},"event":{"type":"quote","text":"AI to be capable of making very small discoveries"},"horizon":{"asStated":"In 2026","words":"In 2026","kind":"absolute","years":[2026],"relation":"in","precision":"year"},"confidence":{"asStated":"\"we expect\"","words":"we expect"},"attested":{"date":"2025-11-07","note":"Internet Archive capture 20251107162450, read by a reviewer from a different model family."}}},{"prediction":{"form":"stated-forecast","predictor":{"kind":"organisation","label":"OpenAI","entityIds":["openai"]},"event":{"type":"quote","text":"we will have systems that can make more significant discoveries"},"horizon":{"asStated":"In 2028 and beyond","words":"In 2028 and beyond","kind":"absolute","years":[2028],"relation":"from","precision":"year"},"confidence":{"asStated":"\"we are pretty confident\"","words":"we are pretty confident"},"hedge":"The sentence adds, in parentheses, that the company could of course be wrong and that this is what its research progress appears to indicate (paraphrased here).","attested":{"date":"2025-11-07","note":"Internet Archive capture 20251107162450, read by a reviewer from a different model family."}}}]}},"sources":{"n":0,"fields":{}},"evidence":{"n":0,"fields":{}}}}