{"entities":[{"id":"eliezer-yudkowsky","name":"Eliezer Yudkowsky","type":"person"},{"id":"google","name":"Google","type":"organisation"},{"id":"deepseek","shortName":"DeepSeek","type":"organisation"},{"id":"deepseek-r1-report","name":"DeepSeek-R1 research report","type":"document"},{"id":"minimax","name":"MiniMax","type":"organisation"},{"id":"lw-2025-re-recent-anthropic-safety-research","shortName":"Yudkowsky, Re: recent Anthropic safety research (2025)","type":"document"},{"id":"jones-2025-spend-to-reduce-ai-existential-risk","shortName":"Jones, How Much Should We Spend to Reduce A.I.'s Existential Risk (2025)","type":"document"},{"id":"agrawal-gans-goldfarb-2025-genius-on-demand","shortName":"Agrawal, Gans and Goldfarb, Genius on Demand (2025)","type":"document"},{"id":"restrepo-2025-we-wont-be-missed","shortName":"Restrepo, We Won't be Missed (2025)","type":"document"},{"id":"andrews-farboodi-2025-markets-transformative-ai","shortName":"Andrews and Farboodi, Do Markets Believe in Transformative AI (2025)","type":"document"},{"id":"brynjolfsson-korinek-agrawal-2025-research-agenda-transformative-ai","shortName":"Brynjolfsson, Korinek and Agrawal, A Research Agenda for the Economics of Transformative AI (2025)","type":"document"},{"id":"baidu","name":"Baidu / 百度","type":"organisation"},{"id":"bytedance-seed","name":"ByteDance Seed / 字节跳动 Seed","type":"organisation"},{"id":"moonshot-ai","name":"Moonshot AI / 月之暗面","type":"organisation"},{"id":"wu-yuxin","name":"Wu Yuxin","type":"person"},{"id":"yang-zhilin","name":"Yang Zhilin","type":"person"},{"id":"zhou-xinyu","name":"Zhou Xinyu","type":"person"},{"id":"ajay-agrawal","name":"Ajay Agrawal","type":"person"},{"id":"erik-brynjolfsson","name":"Erik Brynjolfsson","type":"person"},{"id":"joshua-s-gans","name":"Joshua S. Gans","type":"person"},{"id":"avi-goldfarb","name":"Avi Goldfarb","type":"person"},{"id":"charles-i-jones","name":"Charles I. Jones","type":"person"},{"id":"anton-korinek","name":"Anton Korinek","type":"person"},{"id":"pascual-restrepo","name":"Pascual Restrepo","type":"person"},{"id":"minimax-prospectus-2025","name":"MiniMax Group Inc. Global Offering prospectus (31 December 2025)","type":"document"},{"id":"sail-frontier-ai-risk-management-framework-v1-0","shortName":"Frontier AI Risk Management Framework · v1.0","type":"document","documentVersion":{"family":"sail-frontier-ai-risk-management-framework","label":"v1.0","publisherId":"shanghai-ai-lab","publishedAt":"2025-07","dateNote":"The document's citation line gives July 2025; no day is stated in the inspected text. A Chinese edition and a later v1.5 technical report exist but were not inspected."}},{"id":"suncatcher-blog-2025","shortName":"Google Research, Project Suncatcher post (2025)","type":"document"},{"id":"suncatcher-paper-2025","shortName":"Agüera y Arcas et al., Suncatcher paper (2025)","type":"document"},{"id":"nvidia-starcloud-blog-2025","shortName":"NVIDIA blog on Starcloud (2025)","type":"document"},{"id":"axiom-odc-release-2025","shortName":"Axiom Space, orbital data-centre nodes (2025)","type":"document"},{"id":"espi-report-98-2025","shortName":"ESPI Report 98 (2025)","type":"document"},{"id":"aili-2025-carbon-neutral-data-centres-in-space","shortName":"Aili, Choi, Ong and Wen, Carbon-neutral data centres in space (2025)","type":"document"},{"id":"anthropic-llms-biorisk-2025-09","name":"LLMs and biorisk","type":"document","documentVersion":{"family":"anthropic-llms-biorisk","label":"Anthropic Frontier Red Team post (5 September 2025)","publisherId":"anthropic","publishedAt":"2025-09-05","dateNote":"The page's own metadata reads \"Published Sep 5, 2025\" and names the Frontier Red Team (`anthropic-biorisk`)."}},{"id":"brent-mckelvey-foundation-models-2025-06","name":"Contemporary AI foundation models increase biological weapons risk","type":"document","documentVersion":{"family":"brent-mckelvey-foundation-models","label":"RAND working paper WR-A3853-1, arXiv v1 (12 June 2025)","publisherId":"rand-corporation","publishedAt":"2025-06-12","dateNote":"The arXiv abstract page for 2506.13798 records the v1 submission as 12 June 2025 and the report number WR-A3853-1 (`arxiv-2506-13798-abs`); the PDF's cover reads \"June 2025\" (`brent-mckelvey-2025`). A later RAND perspective (PEA3853-1) with a similar title was seen in search results and was not retrieved."}}],"claims":[{"id":"cn-deepseek-r1","subject":"deepseek","predicate":"published","object":{"kind":"entity","entityId":"deepseek-r1-report"},"topic":"ideas","statement":"DeepSeek published the R1 research report.","time":{"kind":"event","date":"2025-01-22","basis":"publication","note":"Date of this source expression or announcement, not a separately verified implementation date."},"derivation":"explicit","resolution":"supported"},{"id":"cn-r1-rl","subject":"deepseek-r1-report","predicate":"discusses","object":{"kind":"entity","entityId":"reinforcement-learning"},"topic":"ideas","statement":"The R1 report distinguishes R1-Zero’s RL without preliminary SFT from R1’s cold-start training.","time":{"kind":"event","date":"2025-01-22","basis":"publication","note":"Date of this source expression or announcement, not a separately verified implementation date."},"derivation":"explicit","resolution":"supported"},{"id":"cn-r1-qwen","subject":"deepseek-r1-report","predicate":"cites","object":{"kind":"entity","entityId":"qwen25-announcement"},"topic":"ideas","statement":"R1’s distillation section cites Qwen2.5’s announcement while naming Qwen base models fine-tuned on R1-generated samples.","time":{"kind":"event","date":"2025-01-22","basis":"publication","note":"Date of this source expression or announcement, not a separately verified implementation date."},"derivation":"explicit","resolution":"supported"},{"id":"yk-y21-lw-2025-re-recent-anthropic-safety-research-c","subject":"lw-2025-re-recent-anthropic-safety-research","predicate":"contests_proposition","object":{"kind":"entity","entityId":"human-data-humanlike-mind"},"topic":"ideas","statement":"The post “Re: recent Anthropic safety research” (2025-08-06) contests Y21. Commenting on reported behaviour of a language model, he distinguishes a learned human-like persona from the system's underlying preferences, which he expects to be alien.","time":{"kind":"event","date":"2025-08-06","basis":"publication","note":"LessWrong postedAt."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["contesting","later-restatement"],"register":"nonfiction","claimed":{"date":"2025-08-04","note":"Decoded from the linked X status identifier; the X original was not retrieved and the date is unverified (cross-family check, 2026-10-07)."},"engagement":{"kind":"contrary-only","rationale":"Under the message-level rule, the passage states the contrary conclusion without giving a reason against the premise that training on human text or behaviour yields human-like thought and motives, even when read with its message or dialogue turn."}}}},{"id":"ec-e02-jones-2025-spend-to-reduce-ai-existential-risk-b","subject":"jones-2025-spend-to-reduce-ai-existential-risk","predicate":"bears_on_proposition","object":{"kind":"entity","entityId":"machine-intelligence-exceeds-humans-in-all-domains"},"topic":"ideas","statement":"Jones, “How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) bears on E02. The preceding sentence reports experts on catastrophic outcomes 'either from malicious use or even from a superintelligent A.I. itself'.","time":{"kind":"event","date":"2025-03","basis":"publication","note":"Cover \"March 2025\". Checked in a current copy; an archive capture was not available, so earlier bytes are not authenticated."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"'Most tasks', not essentially all; framed as world-changing benefit.","tier":{"value":1,"basis":"NBER working paper (core venue); Jones is an economist."}}}},{"id":"ec-e02-agrawal-gans-goldfarb-2025-genius-on-demand-b","subject":"agrawal-gans-goldfarb-2025-genius-on-demand","predicate":"bears_on_proposition","object":{"kind":"entity","entityId":"machine-intelligence-exceeds-humans-in-all-domains"},"topic":"ideas","statement":"Agrawal, Gans and Goldfarb, “Genius on Demand: The Value of Transformative Artificial Intelligence” (2025-10) bears on E02. Model of 'genius on demand' AI; human geniuses specialise where their comparative advantage over AI is greatest.","time":{"kind":"event","date":"2025-10","basis":"publication","note":"Crossref issued 2025-10. PDF refused (HTTP 403); NBER abstract page used. Checked in a current copy; an archive capture was not available, so earlier bytes are not authenticated."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["context"],"register":"nonfiction","qualifier":"A model scenario under stated conditions, not the author’s own claim.","missing":"AI approaching, not exceeding, human genius efficiency; in knowledge work only.","tier":{"value":1,"basis":"NBER working paper (core venue); Agrawal, Gans and Goldfarb are economists."},"conditions":"'if AI efficiency approaches human genius efficiency'."}}},{"id":"ec-e02-restrepo-2025-we-wont-be-missed-b","subject":"restrepo-2025-we-wont-be-missed","predicate":"bears_on_proposition","object":{"kind":"entity","entityId":"machine-intelligence-exceeds-humans-in-all-domains"},"topic":"ideas","statement":"Restrepo, “We Won't be Missed: Work and Growth in the AGI World” (2025-10) bears on E02. Restrepo's chapter for the NBER volume The Economics of Transformative AI; the abstract derives labour-share convergence to zero.","time":{"kind":"event","date":"2025-10","basis":"publication","note":"Crossref issued 2025-10. PDF refused (HTTP 403); NBER abstract page used. Checked in a current copy; an archive capture was not available, so earlier bytes are not authenticated."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"All economically valuable work done by compute (M05); no statement that machine cognition surpasses human cognition.","tier":{"value":1,"basis":"NBER working paper (core venue); Restrepo is an economist."}}}},{"id":"ec-e03-jones-2025-spend-to-reduce-ai-existential-risk-a","subject":"jones-2025-spend-to-reduce-ai-existential-risk","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"ai-existential-risk-in-economic-analysis"},"topic":"ideas","statement":"Jones, “How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) articulates E03. Sentence continues 'in the Covid-19 pandemic'; Jones adds that mortality risk to people alive 'more than justifies spending large amounts'. Abstract follows a back-of-envelope calculation from a $10 million value of life and a 1% mortality risk. Preceded by the note that 12–33% of runs lead to zero mitigation investment (very small or hard-to-mitigate risk). After saying that valuing future generations would raise the amount.","time":{"kind":"event","date":"2025-03","basis":"publication","note":"Cover \"March 2025\". Checked in a current copy; an archive capture was not available, so earlier bytes are not authenticated."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","qualifier":"A model scenario under stated conditions, not the author’s own claim.","tier":{"value":1,"basis":"NBER working paper (core venue); Jones is an economist."},"conditions":"(E03-18) 'For most of the scenarios and parameter combinations considered'; no weight on future generations. (E03-19) Monte Carlo average over assumed parameter distributions; zero altruism to future generations."}}},{"id":"ec-e04-jones-2025-spend-to-reduce-ai-existential-risk-a","subject":"jones-2025-spend-to-reduce-ai-existential-risk","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"ai-developer-incentives-and-safety"},"topic":"ideas","statement":"Jones, “How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) articulates E04. Listed among 'many related questions that merit further study'.","time":{"kind":"event","date":"2025-03","basis":"publication","note":"Cover \"March 2025\". Checked in a current copy; an archive capture was not available, so earlier bytes are not authenticated."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","tier":{"value":1,"basis":"NBER working paper (core venue); Jones is an economist."}}}},{"id":"ec-e05-jones-2025-spend-to-reduce-ai-existential-risk-a","subject":"jones-2025-spend-to-reduce-ai-existential-risk","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"economics-engagement-with-superintelligence"},"topic":"ideas","statement":"Jones, “How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) articulates E05. Jones then says the Covid-19 experience supplies a key insight, letting existential risk be modelled like pandemic mortality.","time":{"kind":"event","date":"2025-03","basis":"publication","note":"Cover \"March 2025\". Checked in a current copy; an archive capture was not available, so earlier bytes are not authenticated."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","tier":{"value":1,"basis":"NBER working paper (core venue); Jones is an economist."}}}},{"id":"ec-e05-jones-2025-spend-to-reduce-ai-existential-risk-b","subject":"jones-2025-spend-to-reduce-ai-existential-risk","predicate":"bears_on_proposition","object":{"kind":"entity","entityId":"economics-engagement-with-superintelligence"},"topic":"ideas","statement":"Jones, “How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) bears on E05. Related-literature paragraph that also cites Aschenbrenner (2020), Aschenbrenner and Trammell (2024) and Martin and Pindyck.","time":{"kind":"event","date":"2025-03","basis":"publication","note":"Cover \"March 2025\". Checked in a current copy; an archive capture was not available, so earlier bytes are not authenticated."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"Says much of the work on AI existential risk lies outside economics; it is a remark on a review, not a statement that economics neglected the topic.","tier":{"value":1,"basis":"NBER working paper (core venue); Jones is an economist."}}}},{"id":"ec-e05-andrews-farboodi-2025-markets-transformative-ai-b","subject":"andrews-farboodi-2025-markets-transformative-ai","predicate":"bears_on_proposition","object":{"kind":"entity","entityId":"economics-engagement-with-superintelligence"},"topic":"ideas","statement":"Andrews and Farboodi, “Do Markets Believe in Transformative AI?” (2025-09) bears on E05. Event study of US Treasury and TIPS yields around 30 frontier-lab model releases (Nov 2022–Dec 2025).","time":{"kind":"event","date":"2025-09","basis":"publication","note":"Crossref issued 2025-09. PDF refused (HTTP 403); NBER abstract page used. Checked in a current copy; an archive capture was not available, so earlier bytes are not authenticated."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"Evidence about investors' beliefs on transformative AI, inferred from yields; not about superintelligence, self-improvement or existential risk, and not a statement about economics as a field.","attested":{"date":"2026-07","note":"The NBER abstract page shows \"Revision Date July 2026\"; the abstract describes data to December 2025."},"tier":{"value":1,"basis":"NBER working paper (core venue); Andrews and Farboodi are economists."}}}},{"id":"ec-e05-brynjolfsson-korinek-agrawal-2025-research-agenda-transformative-ai-b","subject":"brynjolfsson-korinek-agrawal-2025-research-agenda-transformative-ai","predicate":"bears_on_proposition","object":{"kind":"entity","entityId":"economics-engagement-with-superintelligence"},"topic":"ideas","statement":"Brynjolfsson, Korinek and Agrawal, “A Research Agenda for the Economics of Transformative AI” (2025-09) bears on E05. Opening sentence of the research-agenda abstract ('As we approach Transformative Artificial Intelligence (TAI), …'). The agenda's list; 'safety risks' is one of nine challenges.","time":{"kind":"event","date":"2025-09","basis":"publication","note":"Crossref issued 2025-09. PDF refused (HTTP 403); NBER abstract page used. Checked in a current copy; an archive capture was not available, so earlier bytes are not authenticated."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"(E05-29) Calls for an economics of transformative AI in general; the abstract does not name superintelligence, AGI, the singularity or existential risk. (E05-30) \"Safety risks\" in an abstract list; existential risk is not stated in the retrieved text.","tier":{"value":1,"basis":"NBER working paper (core venue); Brynjolfsson, Korinek and Agrawal are economists."}}}},{"id":"ap-ajay-agrawal-brynjolfsson-korinek-agrawal-2025-research-agenda-transformative-ai","subject":"ajay-agrawal","predicate":"authored","object":{"kind":"entity","entityId":"brynjolfsson-korinek-agrawal-2025-research-agenda-transformative-ai"},"topic":"ideas","statement":"“A Research Agenda for the Economics of Transformative AI” (2025-09) credits Ajay K. Agrawal as one of its 3 authors.","time":{"kind":"event","date":"2025-09","basis":"publication","note":"Date the source record gives for this text; authorship is credited in the inspected expression."},"derivation":"explicit","resolution":"supported"},{"id":"ap-ajay-agrawal-agrawal-gans-goldfarb-2025-genius-on-demand","subject":"ajay-agrawal","predicate":"authored","object":{"kind":"entity","entityId":"agrawal-gans-goldfarb-2025-genius-on-demand"},"topic":"ideas","statement":"“Genius on Demand: The Value of Transformative Artificial Intelligence” (2025-10) credits Ajay K. Agrawal as one of its 3 authors.","time":{"kind":"event","date":"2025-10","basis":"publication","note":"Date the source record gives for this text; authorship is credited in the inspected expression."},"derivation":"explicit","resolution":"supported"},{"id":"ap-erik-brynjolfsson-brynjolfsson-korinek-agrawal-2025-research-agenda-transformative-ai","subject":"erik-brynjolfsson","predicate":"authored","object":{"kind":"entity","entityId":"brynjolfsson-korinek-agrawal-2025-research-agenda-transformative-ai"},"topic":"ideas","statement":"“A Research Agenda for the Economics of Transformative AI” (2025-09) credits Erik Brynjolfsson as one of its 3 authors.","time":{"kind":"event","date":"2025-09","basis":"publication","note":"Date the source record gives for this text; authorship is credited in the inspected expression."},"derivation":"explicit","resolution":"supported"},{"id":"ap-joshua-s-gans-agrawal-gans-goldfarb-2025-genius-on-demand","subject":"joshua-s-gans","predicate":"authored","object":{"kind":"entity","entityId":"agrawal-gans-goldfarb-2025-genius-on-demand"},"topic":"ideas","statement":"“Genius on Demand: The Value of Transformative Artificial Intelligence” (2025-10) credits Joshua S. Gans as one of its 3 authors.","time":{"kind":"event","date":"2025-10","basis":"publication","note":"Date the source record gives for this text; authorship is credited in the inspected expression."},"derivation":"explicit","resolution":"supported"},{"id":"ap-avi-goldfarb-agrawal-gans-goldfarb-2025-genius-on-demand","subject":"avi-goldfarb","predicate":"authored","object":{"kind":"entity","entityId":"agrawal-gans-goldfarb-2025-genius-on-demand"},"topic":"ideas","statement":"“Genius on Demand: The Value of Transformative Artificial Intelligence” (2025-10) credits Avi Goldfarb as one of its 3 authors.","time":{"kind":"event","date":"2025-10","basis":"publication","note":"Date the source record gives for this text; authorship is credited in the inspected expression."},"derivation":"explicit","resolution":"supported"},{"id":"ap-charles-i-jones-jones-2025-spend-to-reduce-ai-existential-risk","subject":"charles-i-jones","predicate":"authored","object":{"kind":"entity","entityId":"jones-2025-spend-to-reduce-ai-existential-risk"},"topic":"ideas","statement":"“How Much Should We Spend to Reduce A.I.'s Existential Risk?” (2025-03) credits Charles I. Jones as its author.","time":{"kind":"event","date":"2025-03","basis":"publication","note":"Date the source record gives for this text; authorship is credited in the inspected expression."},"derivation":"explicit","resolution":"supported"},{"id":"ap-anton-korinek-brynjolfsson-korinek-agrawal-2025-research-agenda-transformative-ai","subject":"anton-korinek","predicate":"authored","object":{"kind":"entity","entityId":"brynjolfsson-korinek-agrawal-2025-research-agenda-transformative-ai"},"topic":"ideas","statement":"“A Research Agenda for the Economics of Transformative AI” (2025-09) credits Anton Korinek as one of its 3 authors.","time":{"kind":"event","date":"2025-09","basis":"publication","note":"Date the source record gives for this text; authorship is credited in the inspected expression."},"derivation":"explicit","resolution":"supported"},{"id":"ap-pascual-restrepo-restrepo-2025-we-wont-be-missed","subject":"pascual-restrepo","predicate":"authored","object":{"kind":"entity","entityId":"restrepo-2025-we-wont-be-missed"},"topic":"ideas","statement":"“We Won't be Missed: Work and Growth in the AGI World” (2025-10) credits Pascual Restrepo as its author.","time":{"kind":"event","date":"2025-10","basis":"publication","note":"Date the source record gives for this text; authorship is credited in the inspected expression."},"derivation":"explicit","resolution":"supported"},{"id":"ap-eliezer-yudkowsky-lw-2025-re-recent-anthropic-safety-research","subject":"eliezer-yudkowsky","predicate":"authored","object":{"kind":"entity","entityId":"lw-2025-re-recent-anthropic-safety-research"},"topic":"ideas","statement":"“Re: recent Anthropic safety research” (2025-08-06) credits Eliezer Yudkowsky as its author.","time":{"kind":"event","date":"2025-08-06","basis":"publication","note":"Date the source record gives for this text; authorship is credited in the inspected expression."},"derivation":"explicit","resolution":"supported"},{"id":"c2-minimax-01-report","subject":"minimax","predicate":"published","object":{"kind":"entity","entityId":"minimax-01-report"},"topic":"ideas","statement":"MiniMax posted the MiniMax-01 technical report, “MiniMax-01: Scaling Foundation Models with Lightning Attention”, to arXiv on 14 January 2025 (version 1).","time":{"kind":"event","date":"2025-01-14","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions."},"derivation":"explicit","resolution":"supported"},{"id":"c2-minimax-01-readme","subject":"minimax","predicate":"published","object":{"kind":"entity","entityId":"minimax-01-readme"},"topic":"ideas","statement":"MiniMax’s MiniMax-01 GitHub repository README, at a commit of 15 January 2025, introduced two models, MiniMax-Text-01 and MiniMax-VL-01.","time":{"kind":"event","date":"2025-01-15","basis":"publication","note":"Date of the cited README commit; the repository’s initial commit is dated 14 January 2025. The prospectus dates the launch of MiniMax-Text-01 to January 2025."},"derivation":"explicit","resolution":"supported"},{"id":"c2-bytedance-seed-doubao-15-pro","subject":"bytedance-seed","predicate":"published","object":{"kind":"entity","entityId":"doubao-15-pro-page"},"topic":"ideas","statement":"A Chinese-language page on ByteDance Seed's website, dated 22 January 2025 and titled 'Doubao 1.5pro - Doubao Team', describes the Doubao-1.5-pro model as using a mixture-of-experts architecture.","time":{"kind":"event","date":"2025-01-22","basis":"publication","note":"Date printed on the page; the page is mutable and was read on 27 October 2026, so the date and wording as originally published are not authenticated."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k15-api-access","subject":"moonshot-ai","predicate":"published","object":{"kind":"entity","entityId":"kimi-k15-readme"},"topic":"ideas","statement":"The Kimi-k1.5 repository README, as of 22 January 2025, offered testing of Kimi k1.5 through the Kimi OpenPlatform API on application, and published no model weights.","time":{"kind":"event","date":"2025-01-22","basis":"publication","note":"Date of the cited README commit; the repository was initialised on 20 January 2025; access terms may have changed later."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k15-author-wu","subject":"wu-yuxin","predicate":"authored","object":{"kind":"entity","entityId":"kimi-k15-report"},"topic":"ideas","statement":"Yuxin Wu is listed among the authors of the Kimi k1.5 technical report (arXiv version 1).","time":{"kind":"event","date":"2025-01-22","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k15-author-yang","subject":"yang-zhilin","predicate":"authored","object":{"kind":"entity","entityId":"kimi-k15-report"},"topic":"ideas","statement":"Zhilin Yang is listed among the authors of the Kimi k1.5 technical report (arXiv version 1).","time":{"kind":"event","date":"2025-01-22","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k15-author-zhou","subject":"zhou-xinyu","predicate":"authored","object":{"kind":"entity","entityId":"kimi-k15-report"},"topic":"ideas","statement":"Xinyu Zhou is listed among the authors of the Kimi k1.5 technical report (arXiv version 1).","time":{"kind":"event","date":"2025-01-22","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k15-report","subject":"moonshot-ai","predicate":"published","object":{"kind":"entity","entityId":"kimi-k15-report"},"topic":"ideas","statement":"Moonshot AI’s Kimi Team posted the Kimi k1.5 technical report to arXiv on 22 January 2025 (version 1).","time":{"kind":"event","date":"2025-01-22","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions."},"derivation":"explicit","resolution":"supported"},{"id":"c2-baidu-ernie-45-x1-launch","subject":"baidu","predicate":"published","object":{"kind":"entity","entityId":"ernie-45-x1-launch-2025"},"topic":"ideas","statement":"Baidu announced ERNIE 4.5 and the reasoning model ERNIE X1 on 16 March 2025, stating that both were accessible to individual users free of charge through ERNIE Bot.","time":{"kind":"event","date":"2025-03-16","basis":"announcement","note":"Dateline and distribution date of the press release. The release says ERNIE X1 was 'set to be available' on the Qianfan platform for enterprise users at a later date."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-kimi-vl-card","subject":"moonshot-ai","predicate":"published","object":{"kind":"entity","entityId":"kimi-vl-a3b-instruct-card"},"topic":"ideas","statement":"Moonshot AI published the Kimi-VL-A3B-Instruct model on its moonshotai Hugging Face account, with a model card whose metadata gives the licence as “mit”, at a commit of 9 April 2025.","time":{"kind":"event","date":"2025-04-09","basis":"publication","note":"Date of the cited Hugging Face commit; repository commit dates do not establish when the repository became public."},"derivation":"explicit","resolution":"supported"},{"id":"c2-bytedance-seed-thinking-v15-report","subject":"bytedance-seed","predicate":"published","object":{"kind":"entity","entityId":"seed-thinking-v15-report"},"topic":"ideas","statement":"ByteDance Seed posted the Seed-Thinking-v1.5 reasoning-model report to arXiv on 10 April 2025 (version 1); later arXiv versions are titled Seed1.5-Thinking.","time":{"kind":"event","date":"2025-04-10","basis":"publication","note":"Submission date of arXiv version 1; not a model release or API availability date. Version 3 is dated 29 April 2025."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-kimi-vl-report","subject":"moonshot-ai","predicate":"published","object":{"kind":"entity","entityId":"kimi-vl-report"},"topic":"ideas","statement":"Moonshot AI’s Kimi Team posted the Kimi-VL technical report to arXiv on 10 April 2025 (version 1).","time":{"kind":"event","date":"2025-04-10","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions."},"derivation":"explicit","resolution":"supported"},{"id":"c2-bytedance-seed-15-vl-report","subject":"bytedance-seed","predicate":"published","object":{"kind":"entity","entityId":"seed-15-vl-report"},"topic":"ideas","statement":"ByteDance Seed posted the Seed1.5-VL vision-language model technical report to arXiv on 11 May 2025 (version 1).","time":{"kind":"event","date":"2025-05-11","basis":"publication","note":"Submission date of arXiv version 1, from the arXiv stamp and submission history. The same version 1 PDF prints an internal date, “Date: June 13, 2025”, on its first page. The two dates are of different kinds and are kept apart; no explanation of the difference is inferred, and neither is the model’s release or API availability date."},"derivation":"explicit","resolution":"supported"},{"id":"c2-bytedance-seed-coder-report","subject":"bytedance-seed","predicate":"published","object":{"kind":"entity","entityId":"seed-coder-report"},"topic":"ideas","statement":"ByteDance Seed posted the Seed-Coder technical report to arXiv on 4 June 2025 (version 1).","time":{"kind":"event","date":"2025-06-04","basis":"publication","note":"Submission date of arXiv version 1. The Hugging Face repositories for the Seed-Coder-8B models were created on 27 April 2025; the date they became public is not established."},"derivation":"explicit","resolution":"supported"},{"id":"c2-minimax-m1-report","subject":"minimax","predicate":"published","object":{"kind":"entity","entityId":"minimax-m1-report"},"topic":"ideas","statement":"MiniMax posted the MiniMax-M1 technical report to arXiv on 16 June 2025 (version 1); the abstract states that MiniMax publicly releases MiniMax-M1 in versions with 40K and 80K thinking budgets.","time":{"kind":"event","date":"2025-06-16","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions. The prospectus dates the launch of MiniMax-M1 to June 2025."},"derivation":"explicit","resolution":"supported"},{"id":"c2-baidu-ernie-45-repository","subject":"baidu","predicate":"published","object":{"kind":"entity","entityId":"ernie-45-repository-readme"},"topic":"ideas","statement":"The PaddlePaddle/ERNIE repository README at its commit of 30 June 2025 introduces ERNIE 4.5 as a family of ten model variants and states that the models, including pre-trained weights and inference code, have been released on Hugging Face and AI Studio.","time":{"kind":"event","date":"2025-06-30","basis":"revision","note":"Commit date of the cited README revision. Baidu's 20-F filed on 17 March 2026 states that Baidu open-sourced the ERNIE 4.5 series in June 2025; the Hugging Face repositories were created on 28 and 29 June 2025, but the date they became public is not established."},"derivation":"explicit","resolution":"supported"},{"id":"c2-governance-sail-frmf-v1-cites-tc260","subject":"sail-frontier-ai-risk-management-framework-v1-0","predicate":"cites","object":{"kind":"entity","entityId":"tc260-asgf-v1-0"},"topic":"ideas","statement":"The Frontier AI Risk Management Framework (v1.0) states that it builds upon the International AI Safety Report (January 2025) and TC260's AI Safety Governance Framework v1.0, which it cites in a footnote.","time":{"kind":"event","date":"2025-07","basis":"publication","note":"Month of the citing document."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k2-readme","subject":"moonshot-ai","predicate":"published","object":{"kind":"entity","entityId":"kimi-k2-readme"},"topic":"ideas","statement":"The README of Moonshot AI’s Kimi-K2 GitHub repository, at the repository’s initial commit dated 11 July 2025, introduced two model variants, Kimi-K2-Base and Kimi-K2-Instruct, and pointed to their checkpoints on Hugging Face.","time":{"kind":"event","date":"2025-07-11","basis":"publication","note":"Date of the repository’s initial commit, which contains the README (11 July 2025, 14:28 UTC). The cross-family check found that GitHub’s repository record gives a creation time of 3 July 2025; a commit date and a repository creation date are different kinds of date, and neither establishes when the repository or the weights became public. The Hugging Face repository moonshotai/Kimi-K2-Instruct also shows commits from 11 July 2025."},"derivation":"explicit","resolution":"supported"},{"id":"c2-bytedance-seed-x-report","subject":"bytedance-seed","predicate":"published","object":{"kind":"entity","entityId":"seed-x-report"},"topic":"ideas","statement":"ByteDance Seed posted the Seed-X multilingual translation model report to arXiv on 18 July 2025 (version 1).","time":{"kind":"event","date":"2025-07-18","basis":"publication","note":"Submission date of arXiv version 1; the Seed-X Hugging Face repositories were created on 16 July 2025. Neither date is established as the public release date of the weights."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k2-author-wu","subject":"wu-yuxin","predicate":"authored","object":{"kind":"entity","entityId":"kimi-k2-report"},"topic":"ideas","statement":"Yuxin Wu is listed among the authors of the Kimi K2 technical report (arXiv version 1).","time":{"kind":"event","date":"2025-07-28","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k2-author-yang","subject":"yang-zhilin","predicate":"authored","object":{"kind":"entity","entityId":"kimi-k2-report"},"topic":"ideas","statement":"Zhilin Yang is listed among the authors of the Kimi K2 technical report (arXiv version 1).","time":{"kind":"event","date":"2025-07-28","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k2-author-zhou","subject":"zhou-xinyu","predicate":"authored","object":{"kind":"entity","entityId":"kimi-k2-report"},"topic":"ideas","statement":"Xinyu Zhou is listed among the authors of the Kimi K2 technical report (arXiv version 1).","time":{"kind":"event","date":"2025-07-28","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k2-report","subject":"moonshot-ai","predicate":"published","object":{"kind":"entity","entityId":"kimi-k2-report"},"topic":"ideas","statement":"Moonshot AI’s Kimi Team posted the Kimi K2 technical report, “Kimi K2: Open Agentic Intelligence”, to arXiv on 28 July 2025 (version 1).","time":{"kind":"event","date":"2025-07-28","basis":"publication","note":"Submission date of the cited arXiv version; not the model’s release date and not the date of later versions."},"derivation":"explicit","resolution":"supported"},{"id":"c2-bytedance-seed-oss-release","subject":"bytedance-seed","predicate":"published","object":{"kind":"entity","entityId":"seed-oss-model-card"},"topic":"ideas","statement":"ByteDance Seed's Seed-OSS model card states that on 20 August 2025 it released Seed-OSS-36B-Base, in versions with and without synthetic data, and Seed-OSS-36B-Instruct.","time":{"kind":"event","date":"2025-08-20","basis":"announcement","note":"Release date as given in the card's News line, read at the revision of 26 August 2025; the Hugging Face repository's initial commit is also dated 20 August 2025."},"derivation":"explicit","resolution":"supported"},{"id":"c2-minimax-m2-announcement","subject":"minimax","predicate":"published","object":{"kind":"entity","entityId":"minimax-m2-announcement"},"topic":"ideas","statement":"MiniMax announced on 27 October 2025 that it was making MiniMax M2 open source and launching it, and stated that the complete model weights had been released on Hugging Face.","time":{"kind":"event","date":"2025-10-27","basis":"announcement","note":"Date shown on the news page (mutable page, read on 27 October 2026). The prospectus dates the launch of MiniMax-M2 to October 2025."},"derivation":"explicit","resolution":"supported"},{"id":"c2-moonshot-k2-thinking-card","subject":"moonshot-ai","predicate":"published","object":{"kind":"entity","entityId":"kimi-k2-thinking-card"},"topic":"ideas","statement":"Moonshot AI published the Kimi-K2-Thinking model on its moonshotai Hugging Face account; the model card, at a commit of 6 November 2025, describes it as a thinking model built starting from Kimi K2.","time":{"kind":"event","date":"2025-11-06","basis":"publication","note":"Date of the cited Hugging Face commit; the repository’s initial commit is dated 4 November 2025. Not a verified moment of public availability."},"derivation":"explicit","resolution":"supported"},{"id":"c2-minimax-prospectus-abab1","subject":"minimax-prospectus-2025","predicate":"states_provision","object":{"kind":"literal","value":"Earliest business milestone: abab1 launched in April 2022"},"topic":"ideas","time":{"kind":"event","date":"2025-12-31","basis":"publication","note":"Statement in the prospectus of 31 December 2025; the launch is dated to the month as the filing gives it."},"resolution":"supported"},{"id":"cpo-orb-oc1-suncatcher-2025","subject":"suncatcher-blog-2025","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"oc1-orbital-solar-supply"},"topic":"ideas","statement":"Google's Suncatcher post (November 2025) states that in the right orbit a solar panel can be up to eight times more productive than on Earth and produce power nearly continuously.","time":{"kind":"event","date":"2025-11-04","basis":"publication","note":"Date of publication, as the source gives it."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction"}}},{"id":"cpo-orb-oc3-axiom-2025","subject":"axiom-odc-release-2025","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"oc3-terrestrial-constraints"},"topic":"ideas","statement":"Axiom Space's release (April 2025) states that limitations of terrestrial data centres, including energy, cooling, real estate, licensing and permitting, are becoming increasingly apparent, and presents its orbital nodes as mitigating them.","time":{"kind":"event","date":"2025-04-07","basis":"publication","note":"Date of publication, as the source gives it."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction"}}},{"id":"cpo-orb-oc4-suncatcher-paper-2025","subject":"suncatcher-paper-2025","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"oc4-launch-cost-condition"},"topic":"ideas","statement":"The Suncatcher paper (November 2025) argues that if launch to LEO reaches $200 per kilogram, launch cost amortised over spacecraft life could be roughly comparable to terrestrial data-centre energy costs per kW.","time":{"kind":"event","date":"2025-11-22","basis":"publication","note":"Date of publication, as the source gives it."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","estimates":[{"holder":"stated in the text","value":"up to USD 200 per kg","horizon":"as stated in the text","definition":"launch price per kilogram","conditional":true,"conditions":"launch price per kilogram to LEO, stated threshold"}]}}},{"id":"cpo-orb-oc5-nvidia-starcloud-2025","subject":"nvidia-starcloud-blog-2025","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"oc5-lifecycle-emissions"},"topic":"ideas","statement":"NVIDIA's post (October 2025) quotes Starcloud's chief executive Philip Johnston as saying that, apart from launch, orbital data centres would bring tenfold carbon-dioxide savings over their life compared with terrestrial power.","time":{"kind":"event","date":"2025-10-15","basis":"publication","note":"Date of publication, as the source gives it."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction"}}},{"id":"cpo-orb-oc6-suncatcher-blog-2025","subject":"suncatcher-blog-2025","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"oc6-heat-rejection-limit"},"topic":"ideas","statement":"Google's Suncatcher post (November 2025) names thermal management among the significant engineering challenges that remain for space-based ML compute.","time":{"kind":"event","date":"2025-11-04","basis":"publication","note":"Date of publication, as the source gives it."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["earliest-located"],"register":"nonfiction"}}},{"id":"cpo-orb-oc7-suncatcher-paper-2025","subject":"suncatcher-paper-2025","predicate":"bears_on_proposition","object":{"kind":"entity","entityId":"oc7-radiation-effects"},"topic":"ideas","statement":"The Suncatcher paper (November 2025) reports proton-beam tests in which Trillium TPU HBM showed irregularities after 2 krad(Si), about three times the authors' five-year requirement of 750 rad(Si).","time":{"kind":"event","date":"2025-11-22","basis":"publication","note":"Date of publication, as the source gives it."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["context"],"register":"nonfiction","missing":"a statement that orbital radiation requires mitigation: it reports a ground test with a margin over the authors’ requirement"}}},{"id":"cpo-orb-oc8-suncatcher-paper-2025","subject":"suncatcher-paper-2025","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"oc8-no-repair"},"topic":"ideas","statement":"The Suncatcher paper (November 2025) states that replacing failed TPUs by technicians, simple on Earth, is impracticable in space, and that redundant provisioning is the simplest solution.","time":{"kind":"event","date":"2025-11-22","basis":"publication","note":"Date of publication, as the source gives it."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction"}}},{"id":"cpo-orb-oc10-suncatcher-blog-2025","subject":"suncatcher-blog-2025","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"oc10-ground-bandwidth"},"topic":"ideas","statement":"Google's Suncatcher post (November 2025) names high-bandwidth ground communications among the significant engineering challenges that remain.","time":{"kind":"event","date":"2025-11-04","basis":"publication","note":"Date of publication, as the source gives it."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["earliest-located"],"register":"nonfiction"}}},{"id":"cpo-orb-suncatcher-paper-published","subject":"google","predicate":"published","object":{"kind":"entity","entityId":"suncatcher-paper-2025"},"topic":"ideas","statement":"Google researchers posted \"Towards a future space-based, highly scalable AI infrastructure system design\" to arXiv as version 1 on 22 November 2025.","time":{"kind":"event","date":"2025-11-22","basis":"publication","note":"arXiv v1 stamp; the blog announced a preprint on 4 November 2025, in a copy not retrieved; v2 is dated 17 June 2026."},"derivation":"explicit","resolution":"supported"},{"id":"cpo-orb-oc8-espi-2025","subject":"espi-report-98-2025","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"oc8-no-repair"},"topic":"ideas","statement":"ESPI Report 98 (November 2025) states that in-orbit servicing remains so costly that deorbiting a failed satellite and launching a replacement is generally considered more cost-effective than sending crews for repairs, and that the maintenance cycle of a space-based data centre is expected to depend heavily on redundancy, which itself increases costs.","time":{"kind":"event","date":"2025-11-19","basis":"publication","note":"Date of publication on the European Space Policy Institute’s report page; the PDF prints “Published: November 2025”."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["earliest-located"],"register":"nonfiction"}}},{"id":"cpo-orb-oc9-espi-2025","subject":"espi-report-98-2025","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"oc9-launch-cost-too-high"},"topic":"ideas","statement":"ESPI Report 98 (November 2025) models gigawatt-scale space-based data centres against a terrestrial benchmark and finds that, under its modelling assumptions, a constellation architecture has double the benchmark’s capital cost and more than triple its operating cost, and a monolithic architecture an operating cost 38% higher.","time":{"kind":"event","date":"2025-11-19","basis":"publication","note":"Date of publication on the European Space Policy Institute’s report page; the PDF prints “Published: November 2025”."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["earliest-located"],"register":"nonfiction","conditions":"The report’s cost model of gigawatt-scale architectures, with a two-year IT-hardware lifetime; its cost equations give expected values without a measure of uncertainty."}}},{"id":"cpo-orb-oc11-aili-2025","subject":"aili-2025-carbon-neutral-data-centres-in-space","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"oc11-sky-and-debris"},"topic":"ideas","statement":"Aili, Choi, Ong and Wen’s Perspective on carbon-neutral data centres in space (Nature Electronics, 2025) states that retired orbital hardware ends up as space junk that could burn up in the atmosphere, releasing emissions and other harmful chemicals, a disadvantage compared with terrestrial data centres.","time":{"kind":"event","date":"2025-10-27","basis":"publication","note":"Online publication date in the publisher’s metadata, as read by the cross-family check on 2026-10-29; this admission did not read the publisher’s page. The proof read here prints “Received: 21 November 2023” and “Accepted: 7 August 2025”."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["earliest-located"],"register":"nonfiction","attested":{"date":"2025-11-25","note":"Internet Archive capture of 25 November 2025 of the author-hosted proof, which the cross-family check found byte-identical to the copy retrieved at admission; this admission could not reach the Archive."}}}},{"id":"ab-arg-d1-brent-articulates","subject":"brent-mckelvey-foundation-models-2025-06","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"ab-prop-current-system-uplift"},"topic":"ideas","statement":"The RAND working paper \"Contemporary AI foundation models increase biological weapons risk\" (WR-A3853-1, arXiv v1 of 12 June 2025) states that contemporary AI models can meaningfully contribute to biological weapons development, against published safety assessments that assert minimal risk.","time":{"kind":"event","date":"2025-06-12","basis":"publication","note":"Date of the arXiv v1 submission of the working paper, whose cover reads June 2025."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","qualifier":"Stated as a finding of the paper's own model dialogues, which are not described here."}}},{"id":"ab-arg-d2-brent-contests","subject":"brent-mckelvey-foundation-models-2025-06","predicate":"contests_proposition","object":{"kind":"entity","entityId":"ab-prop-physical-bottleneck"},"topic":"ideas","statement":"The RAND working paper of 12 June 2025 states that safety assessments mistakenly assume that biological weapons development requires tacit knowledge, that is, expertise gained through experience that cannot be conveyed in words.","time":{"kind":"event","date":"2025-06-12","basis":"publication","note":"Date of the arXiv v1 submission of the working paper, attested by the arXiv listing page; the paper's own cover gives only June 2025."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["contesting"],"register":"nonfiction","qualifier":"Contests the premise as the paper states other assessments hold it.","engagement":{"kind":"engages-premise","rationale":"The paper names the tacit-knowledge premise and argues against it from historical cases."}}}},{"id":"ab-arg-d3-brent-articulates","subject":"brent-mckelvey-foundation-models-2025-06","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"ab-prop-evaluation-validity"},"topic":"ideas","statement":"The RAND working paper of 12 June 2025 states that flawed safety assessments resting on tacit knowledge and inadequate benchmarks may create a false sense of security, leading to weak safeguards and regulatory gaps.","time":{"kind":"event","date":"2025-06-12","basis":"publication","note":"Date of the arXiv v1 submission of the working paper."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","qualifier":"Argues that the direction of the measurement error is an understatement of risk."}}},{"id":"ab-arg-d7-brent-articulates","subject":"brent-mckelvey-foundation-models-2025-06","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"ab-prop-skilled-actor-scope"},"topic":"ideas","statement":"The RAND working paper of 12 June 2025 states that safety assessments rely on benchmarks that overlook the full range of potential threat actors and fail to account for how AI could enhance the capabilities of skilled individuals as well as enabling non-experts.","time":{"kind":"event","date":"2025-06-12","basis":"publication","note":"Date of the arXiv v1 submission of the working paper, attested by the arXiv listing page; the paper's own cover gives only June 2025."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["earliest-located"],"register":"nonfiction","qualifier":"States the point as a criticism of benchmark coverage."}}},{"id":"ab-arg-d2-anthropic-contests","subject":"anthropic-llms-biorisk-2025-09","predicate":"contests_proposition","object":{"kind":"entity","entityId":"ab-prop-physical-bottleneck"},"topic":"ideas","statement":"Anthropic's Frontier Red Team post \"LLMs and biorisk\" (5 September 2025) states that its 2024 pilot study, while not providing strong evidence either way about AI uplift, suggested that tacit knowledge may not be as significant a bottleneck as many have suggested.","time":{"kind":"event","date":"2025-09-05","basis":"publication","note":"Publication date given on the post."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["contesting"],"register":"nonfiction","qualifier":"A developer's own post; the suggestion is hedged in the source.","engagement":{"kind":"engages-premise","rationale":"The post gives its own pilot study as a reason to doubt the premise, while stating the study's limits."}}}},{"id":"ab-arg-d3-anthropic-articulates","subject":"anthropic-llms-biorisk-2025-09","predicate":"articulates_proposition","object":{"kind":"entity","entityId":"ab-prop-evaluation-validity"},"topic":"ideas","statement":"Anthropic's Frontier Red Team post of 5 September 2025 states that text-based uplift trials of the kind it reports are imperfect proxies for real-world scenarios, which involve further factors it names.","time":{"kind":"event","date":"2025-09-05","basis":"publication","note":"Publication date given on the post."},"derivation":"explicit","resolution":"supported","attributes":{"thought":{"roles":["articulation"],"register":"nonfiction","qualifier":"The same sentence states that the trials show plausibility; the record does not separate the two halves."}}}],"sources":[],"evidence":[],"at":{"entities":[19,39,172,176,248,436,602,603,604,611,612,635,643,747,825,829,842,850,859,881,883,889,890,912,986,989,1092,1093,1094,1096,1104,1105,1232,1235],"claims":[362,363,364,1132,1514,1515,1516,1529,1540,1561,1562,1563,1564,2185,2186,2199,2236,2245,2272,2278,2312,2472,2767,2772,2773,2774,2775,2776,2777,2778,2779,2780,2781,2782,2783,2784,2787,2789,2790,2797,2798,2800,2801,2802,2803,2805,2815,2817,2827,3079,3085,3088,3090,3093,3097,3100,3105,3112,3115,3116,3117,3431,3432,3433,3434,3435,3436],"sources":[],"evidence":[]},"version":"1lg47t7"}