{ "schema_version": 1, "name": "AI Forecast Ledger", "description": "Dated, falsifiable public predictions about AI, each graded against archived third-party evidence.", "publisher": "Setpat Technologies, Unipessoal Lda.", "license": "CC-BY-4.0", "license_url": "https://creativecommons.org/licenses/by/4.0/", "attribution": "AI Forecast Ledger (https://aiforecastledger.com/), CC BY 4.0", "homepage": "https://aiforecastledger.com/", "json_url": "https://aiforecastledger.com/ledger.json", "csv_url": "https://aiforecastledger.com/ledger.csv", "generated_at": "2026-11-07T15:58:43.051Z", "updated_at": "2026-11-05", "claim_count": 11, "claims": [ { "id": "superhuman-coder-2027", "url": "https://aiforecastledger.com/claims/superhuman-coder-2027.html", "claim": "A superhuman coder exists by end of 2027", "short_title": "Superhuman coder", "predictor": "AI-2027 report", "claimant_type": "report", "predicted_on": "2025-04", "target_date": "2027-12", "category": "capabilities", "grading_rubric": "Met when a model autonomously completes a non-trivial PR end-to-end at senior-eng level", "metric": "SWE-bench Verified %", "verdict": "on-track", "reviewer_status": "graded", "label": "On track", "last_reviewed": "2026-11-04", "current_measurement": { "metric_value": "79.20", "metric_unit": "% resolved on SWE-bench Verified (top agent, bash-only default view)", "as_of": "2026-11-04", "source_quote": "Claude 4.5 Opus | Sonar Foundation Agent | 79.20", "quote_sha256": "aab3aa48de59ea386155a200e1848822a3d6ff65b292fc64b1da2f1dca989be3", "source_url": "https://www.swebench.com/", "archive_url": "https://web.archive.org/web/20261104000328/https://www.swebench.com/" }, "model_verifications": 2, "evidence": [ { "role": "prediction", "source_name": "AI 2027 (Kokotajlo, Alexander, Larsen, Lifland & Dean)", "source_url": "https://ai-2027.com/", "source_date": "2025-04", "archive_url": "https://web.archive.org/web/20260813090511/https://ai-2027.com/", "origin_org": "AI Futures Project", "exact_quote_or_figure": "Why we forecast a superhuman coder in early 2027", "quote_sha256": "bbe56b08be1bca6f889ede38fe183bf4acc0ced9b8e9a68c945da5bf0583f000", "content_sha256": "88a55463eb8b71a12ddbcbb9fbbf0cdebcd1fccc640c74d8f89bc4cd3781afff", "fetched_at": "2026-08-15T08:40:00Z" }, { "role": "measurement", "source_name": "SWE-bench Verified leaderboard", "source_url": "https://www.swebench.com/", "source_date": "2026-07-15", "archive_url": "https://web.archive.org/web/20260715000433/https://www.swebench.com/", "origin_org": "SWE-bench", "exact_quote_or_figure": "Claude 4.5 Opus (high reasoning) | 76.80", "quote_sha256": "ef37a8d545ee220a2ed801304f343fdf3cbdcb45fb0f81f75e94aea239cea7b5", "content_sha256": "d637eef9330385a2006e62ca85a47ca2cbfc1a5e366cabba2ac3a2c962d1c205", "fetched_at": "2026-07-15T00:02:35.944Z" }, { "role": "measurement", "source_name": "METR — Measuring AI Ability to Complete Long Tasks", "source_url": "https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "source_date": "2025-03-19", "archive_url": "https://web.archive.org/web/2/https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "origin_org": "METR", "exact_quote_or_figure": "best AI agents are not currently able to carry out substantive projects by themselves or directly substitute for human labor.", "quote_sha256": "7f355ccaf73089265cf4174aed9675df37a8c604bac6f927286f1ce7a02537e7", "content_sha256": "1265d924a71b1de96aae7daea892f3c2f8b1e7064eb6bb89e7148b73977129ee", "fetched_at": "2026-07-15T00:02:35.243Z" } ], "prediction_primary": null, "history": [ { "date": "2026-06-20", "change": "seeded", "pr": "#1" }, { "date": "2026-06-24", "change": "replaced placeholder evidence with the SWE-bench Verified leaderboard; verdict held on-track", "pr": "" }, { "date": "2026-06-25", "change": "demoted pending bulletproof re-grade", "pr": "" }, { "date": "2026-06-26", "change": "added trajectory checkpoints (AI-2027 milestones, on track) + refreshed SWE-bench measurement to >85%", "pr": "" }, { "date": "2026-07-15", "change": "Target date is 2027-12 (future). SWE-bench Verified shows top agents above 76% (swebench.com, measurement, SWE-bench), while METR (measurement, different org) confirms AI agents still cannot…", "pr": "" }, { "date": "2026-08-15", "change": "Verdict on-track remains correct (target 2027-12 is in the future), but a fresh read of the SWE-bench Verified leaderboard (fetched 2026-08-15) shows the top bash-only entry at 79.20% resolved…", "pr": "" }, { "date": "2026-09-10", "change": "Verdict on-track remains correct (target 2027-12 in the future). A fresh read of the SWE-bench Verified leaderboard (fetched 2026-09-10) confirms the top bash-only entry at 79.20% (Claude 4.5 Opus…", "pr": "" }, { "date": "2026-10-16", "change": "Verdict on-track remains correct (target 2027-12 is in the future). A fresh read of the SWE-bench Verified leaderboard (fetched 2026-10-16) confirms the top bash-only entry is unchanged at 79.20%…", "pr": "" }, { "date": "2026-11-04", "change": "Verdict on-track remains correct (target 2027-12 is in the future). A fresh read of the SWE-bench Verified leaderboard (fetched 2026-11-04) confirms the top bash-only entry is unchanged at 79.20%…", "pr": "" } ] }, { "id": "ai-imo-gold-medal-by-2025", "url": "https://aiforecastledger.com/claims/ai-imo-gold-medal-by-2025.html", "claim": "An AI built before the contest earns a gold medal at the International Mathematical Olympiad by 2025", "short_title": "IMO gold medal", "predictor": "Eliezer Yudkowsky (bet vs. Paul Christiano)", "claimant_type": "person", "predicted_on": "2022-02", "target_date": "2025", "category": "benchmarks", "grading_rubric": "Met when, at an IMO through 2025, an AI built beforehand scores at or above the gold-medal cutoff under competition conditions", "metric": "IMO score vs. the gold cutoff (35/42 in 2025)", "verdict": "ahead", "reviewer_status": "graded", "label": "Hit", "last_reviewed": "2026-10-24", "current_measurement": { "metric_value": "35/42", "metric_unit": "points at IMO 2025, gold-medal standard", "as_of": "2025-07-21", "source_quote": "An advanced version of Gemini Deep Think solved five out of the six IMO problems perfectly, earning 35 total points, and achieving gold-medal level performance.", "quote_sha256": "7f6d252179db64ba1009d394dcd19308b068d7455e0bbceb38bfd5a0573037a8", "source_url": "https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/", "archive_url": "https://web.archive.org/web/2/https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/" }, "model_verifications": 2, "evidence": [ { "role": "measurement", "source_name": "Google DeepMind", "source_url": "https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/", "source_date": "2025-07-21", "archive_url": "https://web.archive.org/web/20260625114027/https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/", "origin_org": "Google DeepMind", "exact_quote_or_figure": "An advanced version of Gemini Deep Think solved five out of the six IMO problems perfectly, earning 35 total points, and achieving gold-medal level performance.", "quote_sha256": "7f6d252179db64ba1009d394dcd19308b068d7455e0bbceb38bfd5a0573037a8", "content_sha256": "adb43a9e5df6586d2c2e77d600439726ad9e4fbc9b230c2743ce8c09204320c0", "fetched_at": "2026-06-25T12:34:10.987Z" }, { "role": "corroboration", "source_name": "OpenAI", "source_url": "https://openai.com/index/first-proof-submissions/", "source_date": "2026-02-20", "archive_url": "https://web.archive.org/web/20260510230753/https://openai.com/index/first-proof-submissions/", "origin_org": "OpenAI", "exact_quote_or_figure": "In July 2025, we reached gold medal-level performance on the International Mathematical Olympiad with a general-purpose reasoning model (35/42 points).", "quote_sha256": "abf123e5e382bc6a2260c44708e85e296b14c7d4371b34a0e224640438363a05", "content_sha256": "dad3e11fe09dde9ead668346cc24e4488a7043cd1e3eb31eb040f54fb17ac0b3", "fetched_at": "2026-06-25T12:34:10.832Z" }, { "role": "prediction", "source_name": "LessWrong — IMO challenge bet with Eliezer (Paul Christiano)", "source_url": "https://www.lesswrong.com/posts/sWLLdG6DWJEy3CH7n/imo-challenge-bet-with-eliezer", "source_date": "2022-02-25", "archive_url": "https://web.archive.org/web/20260627131904/https://www.lesswrong.com/posts/sWLLdG6DWJEy3CH7n/imo-challenge-bet-with-eliezer", "origin_org": "LessWrong", "exact_quote_or_figure": "So I think we have Paul at <8%, Eliezer at >16% for AI made before the IMO is able to get a gold (under time controls etc. of grand challenge) in one of 2022-2025.", "quote_sha256": "ce119b6923365f1b4759ac1ee2614048587835f141c06d053d75a8c7e6cd62b1", "content_sha256": "6aa2a25b122336f3f0902b7f20a1dd366b0c02c3c2a64e78940f85917205549a", "fetched_at": "2026-06-26T13:11:30.392Z" } ], "prediction_primary": null, "history": [ { "date": "2026-06-24", "change": "seeded", "pr": "" }, { "date": "2026-06-25", "change": "demoted pending bulletproof re-grade", "pr": "" }, { "date": "2026-06-25", "change": "The claim targets 2025. Google DeepMind's official blog (source 0) confirms an advanced Gemini Deep Think scored 35/42 at IMO 2025, meeting the gold-medal cutoff. OpenAI's First Proof blog (source 1) ", "pr": "" }, { "date": "2026-06-26", "change": "added the Yudkowsky–Christiano bet-odds source (LessWrong, >16% vs <8%) to the receipt; context only — claim, grade, and the 2-model verification binding are unchanged", "pr": "" }, { "date": "2026-10-24", "change": "Verdict 'ahead'/'met' remains correct (target 2025 passed with gold-standard performance). A fresh read of the DeepMind blog confirms the same 35/42 measurement; refreshing the receipt.", "pr": "" } ] }, { "id": "hinton-stop-training-radiologists-by-2021", "url": "https://aiforecastledger.com/claims/hinton-stop-training-radiologists-by-2021.html", "claim": "Deep learning outperforms radiologists within five years, so we should stop training them now", "short_title": "Radiologists obsolete", "predictor": "Geoffrey Hinton", "claimant_type": "person", "predicted_on": "2016", "target_date": "2021", "category": "labor", "grading_rubric": "Met if AI had displaced human radiologists (falling demand / headcount) by ~2021; failed if the workforce kept growing and AI became a complement", "metric": "US radiologist workforce headcount vs. 2016", "verdict": "wrong", "reviewer_status": "graded", "label": "Missed", "last_reviewed": "2026-11-03", "current_measurement": { "metric_value": "17.3", "metric_unit": "% growth in US radiologist headcount, 2014-2023", "as_of": "2023", "source_quote": "the number of radiologists increased 17.3% (from 30,723 to 36,024)", "quote_sha256": "ac3a38def7ef0033c3aac8097b026dafdfbbfc7dc6adc86af3dd3181fcd9a469", "source_url": "https://www.neimanhpi.org/press-releases/consolidation-patterns-of-practices-with-radiologists-through-2023/", "archive_url": "https://web.archive.org/web/20261103000600/https://www.neimanhpi.org/press-releases/consolidation-patterns-of-practices-with-radiologists-through-2023/" }, "model_verifications": 2, "evidence": [ { "role": "prediction", "source_name": "UAB Reporter (quoting Geoffrey Hinton, 2016)", "source_url": "https://www.uab.edu/reporter/people-of-uab/this-radiologist-is-helping-doctors-see-through-the-hype-to-an-ai-future", "source_date": "2022-12", "archive_url": "https://web.archive.org/web/20260702101305/https://www.uab.edu/reporter/people-of-uab/this-radiologist-is-helping-doctors-see-through-the-hype-to-an-ai-future", "origin_org": "UAB Reporter", "exact_quote_or_figure": "People should stop training radiologists now. It’s just completely obvious within five years deep learning is going to do better than radiologists", "quote_sha256": "d5d178b115067817a097126d84467572635f439c4bbdcecd2144354e64d7cea9", "content_sha256": "59bc70ed605be0c9010a9ad2490dc92824227efed4d4f36ad5af8be005e754be", "fetched_at": "2026-07-02T11:37:24.062Z" }, { "role": "measurement", "source_name": "Harvey L. Neiman Health Policy Institute", "source_url": "https://www.neimanhpi.org/press-releases/consolidation-patterns-of-practices-with-radiologists-through-2023/", "source_date": "2024", "archive_url": "https://web.archive.org/web/20260702093927/https://www.neimanhpi.org/press-releases/consolidation-patterns-of-practices-with-radiologists-through-2023/", "origin_org": "Harvey L. Neiman Health Policy Institute", "exact_quote_or_figure": "the number of radiologists increased 17.3% (from 30,723 to 36,024)", "quote_sha256": "ac3a38def7ef0033c3aac8097b026dafdfbbfc7dc6adc86af3dd3181fcd9a469", "content_sha256": "b5ca79a7fa78936bc551a488e5a10670dc2c1211270d667ad67cef83b75e2cb1", "fetched_at": "2026-07-02T09:39:13.194Z" }, { "role": "corroboration", "source_name": "Nature", "source_url": "https://www.nature.com/articles/s44401-025-00023-6", "source_date": "2025", "archive_url": "https://web.archive.org/web/20260702093929/https://www.nature.com/articles/s44401-025-00023-6", "origin_org": "Nature", "exact_quote_or_figure": "a significant radiology workforce shortage", "quote_sha256": "58f6e91c917acfe952a5ed50f52841d9df62f1337eb85ad59e7a1aa372a5e94c", "content_sha256": "a15bb1df2aa1b06d14a5f17dc03f80602d4f7c41e1c2c3c4df0d1316a5836fd5", "fetched_at": "2026-07-02T09:39:13.988Z" } ], "prediction_primary": { "kind": "recording", "url": "https://www.youtube.com/watch?v=2HMPRXstSvQ", "publisher": "Creative Destruction Lab", "title": "Geoff Hinton: On Radiology", "date": "2016", "archive_url": "https://web.archive.org/web/20261022175204/https://www.youtube.com/watch?v=2HMPRXstSvQ" }, "history": [ { "date": "2026-06-24", "change": "seeded", "pr": "" }, { "date": "2026-06-25", "change": "demoted pending bulletproof re-grade", "pr": "" }, { "date": "2026-07-02", "change": "Target date (2021) long past. Neiman HPI (measurement): radiologist headcount grew 17.3% (30,723 to 36,024) from 2014-2023; Nature (2025) corroborates a persistent shortage. Graded wrong per rubric.", "pr": "" }, { "date": "2026-08-17", "change": "Verdict 'wrong' remains correct: the 2021 target is long past and the radiologist workforce grew rather than collapsed. A fresher Neiman HPI release (Feb 2025, neimanhpi.org) refreshes the workforce…", "pr": "" }, { "date": "2026-09-14", "change": "Verdict 'wrong' remains correct (2021 target long past; the workforce grew rather than collapsed). A fresh read of the Neiman HPI release (fetched 2026-09-14) provides a concrete headcount…", "pr": "" }, { "date": "2026-11-03", "change": "The verdict 'wrong' remains correct (2021 target long past; the workforce grew rather than collapsed). A fresh read of the Neiman HPI release (fetched 2026-11-03) provides the same concrete…", "pr": "" }, { "date": "2026-11-05", "change": "Linked the primary source: the organiser’s recording of the talk (Creative Destruction Lab). The quote text still comes from the press report in the receipts. Verdict unchanged.", "pr": "" } ] }, { "id": "amodei-ai-writes-90pct-of-code-by-2025", "url": "https://aiforecastledger.com/claims/amodei-ai-writes-90pct-of-code-by-2025.html", "claim": "Within three to six months AI will be writing 90% of code, and within a year essentially all of it", "short_title": "AI writes 90% of code", "predictor": "Dario Amodei (Anthropic CEO)", "claimant_type": "person", "predicted_on": "2025-03", "target_date": "2026-03", "category": "capabilities", "grading_rubric": "Met if, broadly across software, AI is generating ~90% of code by late 2025; failed if human-written code stays the clear majority of production software", "metric": "Share of production code generated by AI vs. written by humans", "verdict": "wrong", "reviewer_status": "graded", "label": "Missed", "last_reviewed": "2026-11-03", "current_measurement": { "metric_value": "30.1", "metric_unit": "% of Python functions (US contributors) written by AI, December 2024", "as_of": "2024-12", "source_quote": "By December 2024, AI wrote an estimated 30.1% of Python functions from U.S. contributors", "quote_sha256": "280497af94a744c78c7286adc2fd8cee6e2268029ef4a306fe13c9331eb48b6f", "source_url": "https://arxiv.org/html/2506.08945v1", "archive_url": "https://web.archive.org/web/20261103000717/https://arxiv.org/html/2506.08945v1" }, "model_verifications": 2, "evidence": [ { "role": "prediction", "source_name": "Council on Foreign Relations (transcript: CEO Speaker Series with Dario Amodei, 2025-03-10)", "source_url": "https://www.cfr.org/event/ceo-speaker-series-dario-amodei-anthropic", "source_date": "2025-03-10", "archive_url": "https://web.archive.org/web/20260418003122/https://www.cfr.org/event/ceo-speaker-series-dario-amodei-anthropic", "origin_org": "Council on Foreign Relations", "exact_quote_or_figure": "I think we’ll be there in three to six months—where AI is writing 90 percent of the code. And then in twelve months, we may be in a world where AI is writing essentially all of the code.", "quote_sha256": "9fef2a0e3edbdcbaf3bbe723b450058e9e3a6ea690b7378af637c96d703707f8", "content_sha256": "0ddc75f69dabb81e2c2300b3b4f427854bf7cbc3b1b6fcd389c1fe80e2be6b2e", "fetched_at": "2026-11-05T14:18:55Z" }, { "role": "corroboration", "source_name": "Yahoo Finance (Amodei at the Council on Foreign Relations, March 2025)", "source_url": "https://finance.yahoo.com/news/anthropic-ceo-says-ai-could-193020957.html", "source_date": "2025-03", "archive_url": "https://web.archive.org/web/20260405125809/https://finance.yahoo.com/news/anthropic-ceo-says-ai-could-193020957.html", "origin_org": "Yahoo Finance", "exact_quote_or_figure": "AI could soon write 90% of software code in as few as three to six months, Anthropic CEO Dario Amodei said at a Council on Foreign Relations event on March 10.", "quote_sha256": "c5f17eede16879a21fe958bae32a1d08eb391412a94f9f72a3530439a09ed44a", "content_sha256": "56dd534fdba2b8736ec75d9d8814bed4d46a02c1c614e737cf9691e7ca20a260", "fetched_at": "2026-07-03T13:40:00.000Z" }, { "role": "measurement", "source_name": "arXiv (Daniotti et al.)", "source_url": "https://arxiv.org/html/2506.08945v1", "source_date": "2025-06", "archive_url": "https://web.archive.org/web/20260703000142/https://arxiv.org/html/2506.08945v1", "origin_org": "Complexity Science Hub / Utrecht University", "exact_quote_or_figure": "By December 2024, AI wrote an estimated 30.1% of Python functions from U.S. contributors", "quote_sha256": "280497af94a744c78c7286adc2fd8cee6e2268029ef4a306fe13c9331eb48b6f", "content_sha256": "cc8b2f19b002e85e4977fdbda375bc2a81af0f9c9776201ba1c775606a86bda0", "fetched_at": "2026-07-03T00:01:18.465Z" }, { "role": "corroboration", "source_name": "TechCrunch", "source_url": "https://techcrunch.com/2025/04/29/microsoft-ceo-says-up-to-30-of-the-companys-code-was-written-by-ai/", "source_date": "2025-04-29", "archive_url": "https://web.archive.org/web/20260703000213/https://techcrunch.com/2025/04/29/microsoft-ceo-says-up-to-30-of-the-companys-code-was-written-by-ai/", "origin_org": "TechCrunch", "exact_quote_or_figure": "20% to 30% of code inside the company’s repositories was “written by software” — meaning AI", "quote_sha256": "7456d1798ba20d6df2483a81ab3ebe56d68e86be82598002f912631d95017d33", "content_sha256": "ab7ffc7f56c4c2167e08a0313b49efbda8447eb8df318ba777051803a5ec87a1", "fetched_at": "2026-07-03T00:01:24.343Z" } ], "prediction_primary": { "kind": "transcript", "url": "https://www.cfr.org/event/ceo-speaker-series-dario-amodei-anthropic", "publisher": "Council on Foreign Relations", "title": "CEO Speaker Series with Dario Amodei of Anthropic", "date": "2025-03-10", "archive_url": "https://web.archive.org/web/20260418003122/https://www.cfr.org/event/ceo-speaker-series-dario-amodei-anthropic" }, "history": [ { "date": "2026-06-24", "change": "seeded", "pr": "" }, { "date": "2026-06-25", "change": "demoted pending bulletproof re-grade", "pr": "" }, { "date": "2026-06-26", "change": "added trajectory checkpoints — both 3–6mo and 1-yr milestones missed", "pr": "" }, { "date": "2026-07-03", "change": "The claim's target date (2026-03) is in the past. The prediction was that AI would be writing 90% of code by late 2025/early 2026. Source 7 (arXiv, measurement) shows that by December 2024, AI wrote…", "pr": "" }, { "date": "2026-09-14", "change": "Verdict 'wrong' remains correct (target 2026-03 passed with AI's share of production code far below 90%). A fresh read of the arXiv study supplies the best quantified independent measurement of…", "pr": "" }, { "date": "2026-11-03", "change": "The verdict 'wrong' remains correct (target 2026-03 passed with AI's share of production code far below 90%). A fresh read of the arXiv study supplies the best quantified independent measurement of…", "pr": "" }, { "date": "2026-11-05", "change": "Prediction receipt now quotes CFR’s own transcript of the 2025-03-10 event, verbatim and archived; the Yahoo Finance report is kept as corroboration. Verdict unchanged.", "pr": "" } ] }, { "id": "metaculus-weakly-general-ai-2028", "url": "https://aiforecastledger.com/claims/metaculus-weakly-general-ai-2028.html", "claim": "The first weakly general AI system is publicly announced around 2028", "short_title": "Weakly-general AI", "predictor": "Metaculus community forecast", "claimant_type": "community", "predicted_on": "2020", "target_date": "2028", "category": "timelines", "grading_rubric": "Met if a system meeting the Metaculus weakly-general-AI resolution criteria is publicly announced near the community median", "metric": "Public announcement date vs. the live community median", "verdict": "undetermined", "reviewer_status": "unverified", "label": "Pending", "last_reviewed": "2026-06-24", "current_measurement": "unknown — pending a direct read of the live Metaculus median", "model_verifications": 0, "evidence": [], "prediction_primary": null, "history": [ { "date": "2026-06-24", "change": "seeded; awaiting a verifiable read of the Metaculus median before grading", "pr": "" } ] }, { "id": "musk-ai-smarter-than-any-human-2025", "url": "https://aiforecastledger.com/claims/musk-ai-smarter-than-any-human-2025.html", "claim": "AI smarter than any one human probably around the end of 2025", "short_title": "AI smarter than any human by end of 2025", "predictor": "Elon Musk", "claimant_type": "person", "predicted_on": "2024-04", "target_date": "2025-12", "category": "capabilities", "grading_rubric": "Met if, by end of 2025, a deployed AI system demonstrably exceeds the most capable individual human across the breadth of cognitive tasks (not just benchmarks); failed if leading systems remain below top human experts in substantial domains", "metric": "Broad cognitive capability vs. the single most capable human", "verdict": "wrong", "reviewer_status": "graded", "label": "Missed", "last_reviewed": "2026-11-03", "current_measurement": { "metric_value": "0%", "metric_unit": "Pure LLM score on ARC-AGI-2", "as_of": "2025-03-24", "source_quote": "Pure LLMs score 0% on ARC-AGI-2", "quote_sha256": "918b4558c6114fbd7ed4664473a3b5f63b6ec820c518c2f179d35b5820ea6ece", "source_url": "https://arcprize.org/blog/announcing-arc-agi-2-and-arc-prize-2025", "archive_url": "https://web.archive.org/web/20261103000826/https://arcprize.org/blog/announcing-arc-agi-2-and-arc-prize-2025" }, "model_verifications": 2, "evidence": [ { "role": "prediction", "source_name": "The Guardian", "source_url": "https://www.theguardian.com/technology/2024/apr/09/elon-musk-predicts-superhuman-ai-will-be-smarter-than-people-next-year", "source_date": "2024-04-09", "archive_url": "https://web.archive.org/web/20260729060140/https://www.theguardian.com/technology/2024/apr/09/elon-musk-predicts-superhuman-ai-will-be-smarter-than-people-next-year", "origin_org": "The Guardian", "exact_quote_or_figure": "“My guess is that we’ll have AI that is smarter than any one human probably around the end of next year,” Musk said in a livestreamed interview on his social network X.", "quote_sha256": "daec5a176edf976d7466ca336856940f42ab1c37be92baed59fa7361a0196209", "content_sha256": "ba7b49cc93f7a90ea1bd44908676e75e463ae94e5fec74c85228f99ba422fbc9", "fetched_at": "2026-08-15T08:40:00Z" }, { "role": "measurement", "source_name": "METR — Measuring AI Ability to Complete Long Tasks", "source_url": "https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "source_date": "2025-03-19", "archive_url": "https://web.archive.org/web/20260705000344/https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "origin_org": "METR", "exact_quote_or_figure": "best AI agents are not currently able to carry out substantive projects by themselves or directly substitute for human labor.", "quote_sha256": "7f355ccaf73089265cf4174aed9675df37a8c604bac6f927286f1ce7a02537e7", "content_sha256": "1265d924a71b1de96aae7daea892f3c2f8b1e7064eb6bb89e7148b73977129ee", "fetched_at": "2026-07-05T00:00:37.528Z" }, { "role": "measurement", "source_name": "ARC Prize — ARC-AGI-2 and ARC Prize 2025", "source_url": "https://arcprize.org/blog/announcing-arc-agi-2-and-arc-prize-2025", "source_date": "2025-03-24", "archive_url": "https://web.archive.org/web/20260705000334/https://arcprize.org/blog/announcing-arc-agi-2-and-arc-prize-2025", "origin_org": "ARC Prize", "exact_quote_or_figure": "Pure LLMs score 0% on ARC-AGI-2", "quote_sha256": "918b4558c6114fbd7ed4664473a3b5f63b6ec820c518c2f179d35b5820ea6ece", "content_sha256": "1d6af529d7c448009b902ff92c92ca6d19703f8442e707b44ed9954aad791e9f", "fetched_at": "2026-07-05T00:02:53.477Z" } ], "prediction_primary": { "kind": "recording", "url": "https://shows.acast.com/622618c7057f3400120d15db/66146b9cfcf85000165c9a78", "publisher": "Norges Bank Investment Management (In Good Company with Nicolai Tangen)", "title": "Elon Musk: AI, Space, X, Mars, speed and hardcore (live recording on X Spaces)", "date": "2024-04" }, "history": [ { "date": "2026-07-05", "change": "Target date 2025-12 is past. The claim required AI to exceed the most capable human across the breadth of cognitive tasks by end of 2025. Two recognized-domain measurements show this did not occur…", "pr": "" }, { "date": "2026-11-03", "change": "The verdict 'wrong' remains correct (target 2025-12 passed without AI exceeding the most capable human across cognitive domains). A fresh read of the ARC Prize source confirms a concrete measurement…", "pr": "" }, { "date": "2026-11-05", "change": "Linked the primary source: the recording of the April 2024 X Spaces interview, published by Norges Bank Investment Management. The quote text still comes from the press report in the receipts. Verdict unchanged.", "pr": "" } ] }, { "id": "altman-agents-join-workforce-2025", "url": "https://aiforecastledger.com/claims/altman-agents-join-workforce-2025.html", "claim": "In 2025, the first AI agents join the workforce and materially change the output of companies", "short_title": "AI agents join the workforce in 2025", "predictor": "Sam Altman (OpenAI CEO)", "claimant_type": "person", "predicted_on": "2025-01", "target_date": "2025-12", "category": "economy", "grading_rubric": "Met if by end of 2025 autonomous AI agents are deployed in real companies doing economically material work (documented output changes), beyond copilots and chat assistants; failed if agent deployments remain pilots without material output impact", "metric": "Documented production deployments of autonomous agents with material output impact", "verdict": "wrong", "reviewer_status": "graded", "label": "Missed", "last_reviewed": "2026-10-22", "current_measurement": { "metric_value": "6:1", "metric_unit": "interactive AI agent hours to employee hours at Anthropic (frontier-adopter case study)", "as_of": "2026-08", "source_quote": "By August interactive AI agent hours outnumbered employees’ six to one.", "quote_sha256": "3a1ef747b96796c2674c7edd9fb5a179d447408f6531719b2d28db78bdb3c501", "source_url": "https://www.anthropic.com/research/employment-at-anthropic", "archive_url": "https://web.archive.org/web/20261022000326/https://www.anthropic.com/research/employment-at-anthropic" }, "model_verifications": 2, "evidence": [ { "role": "prediction", "source_name": "Sam Altman — Reflections (blog)", "source_url": "https://blog.samaltman.com/reflections", "source_date": "2025-01", "archive_url": "https://web.archive.org/web/20260813061003/https://blog.samaltman.com/reflections", "origin_org": "Sam Altman", "exact_quote_or_figure": "We are now confident we know how to build AGI as we have traditionally understood it. We believe that, in 2025, we may see the first AI agents “join the workforce” and materially change the output of companies.", "quote_sha256": "4d9e4093181fe40317f4f66083b780ad6653d0dea90d564eccdcd9734e5230d5", "content_sha256": "056217a7b2b06dceacbf460f1831741187322a76a5c709db4cc78b19b95d8def", "fetched_at": "2026-08-15T08:40:00Z" }, { "role": "measurement", "source_name": "METR — Measuring AI Ability to Complete Long Tasks", "source_url": "https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "source_date": "2025-03-19", "archive_url": "https://web.archive.org/web/20260715000643/https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "origin_org": "METR", "exact_quote_or_figure": "best AI agents are not currently able to carry out substantive projects by themselves or directly substitute for human labor.", "quote_sha256": "7f355ccaf73089265cf4174aed9675df37a8c604bac6f927286f1ce7a02537e7", "content_sha256": "1265d924a71b1de96aae7daea892f3c2f8b1e7064eb6bb89e7148b73977129ee", "fetched_at": "2026-07-15T00:02:35.243Z" }, { "role": "measurement", "source_name": "Fortune — MIT NANDA report on GenAI pilots", "source_url": "https://fortune.com/2025/08/18/mit-report-95-percent-generative-ai-pilots-at-companies-failing-cfo/", "source_date": "2025-08-18", "archive_url": "https://web.archive.org/web/20260715000746/https://fortune.com/2025/08/18/mit-report-95-percent-generative-ai-pilots-at-companies-failing-cfo/", "origin_org": "Fortune", "exact_quote_or_figure": "The 95% failure rate for enterprise AI solutions", "quote_sha256": "595cee17491233cad37f2855474b6f651f3102bc22121301426bbc91fb0a085f", "content_sha256": "e249c1417a9b923963c62dc15806c6e4709884e8a3442f5fb0f1bd523bf37dbd", "fetched_at": "2026-07-15T00:03:57.861Z" } ], "prediction_primary": null, "history": [ { "date": "2026-07-15", "change": "Target date 2025-12 is past. Two recognized-domain measurements confirm AI agents did not materially change company output: METR (metr.org, measurement) states agents cannot carry out substantive…", "pr": "" }, { "date": "2026-10-22", "change": "Verdict 'wrong' remains correct: the 2025-12 deadline passed without documented broad agent deployments materially changing company output (METR measured in March 2025 that agents could not…", "pr": "" } ] }, { "id": "kurzweil-turing-test-by-2029", "url": "https://aiforecastledger.com/claims/kurzweil-turing-test-by-2029.html", "claim": "AI reaches human-level intelligence and passes the Turing test by 2029", "short_title": "Turing test passed by 2029", "predictor": "Ray Kurzweil", "claimant_type": "person", "predicted_on": "2024-06", "target_date": "2029", "category": "timelines", "grading_rubric": "Met if before end of 2029 an AI passes a rigorous adversarial Turing test (expert judges, extended sessions) or an equivalent accepted human-level-intelligence demonstration; failed otherwise", "metric": "Rigorous adversarial Turing-test pass", "verdict": "on-track", "reviewer_status": "graded", "label": "On track", "last_reviewed": "2026-11-03", "current_measurement": { "metric_value": "35/42", "metric_unit": "IMO 2025 score (gold-medal standard)", "as_of": "2025-07-21", "source_quote": "solved five out of the six IMO problems perfectly, earning 35 total points", "quote_sha256": "0fee296a1b5542d778c99d64449222275aa79cdde262cd833deef00d33412e7d", "source_url": "https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/", "archive_url": "https://web.archive.org/web/20261103000938/https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/" }, "model_verifications": 2, "evidence": [ { "role": "measurement", "source_name": "Google DeepMind — Advanced Gemini Deep Think at IMO 2025", "source_url": "https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/", "source_date": "2025-07-21", "archive_url": "https://web.archive.org/web/2/https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/", "origin_org": "Google DeepMind", "exact_quote_or_figure": "solved five out of the six IMO problems perfectly, earning 35 total points", "quote_sha256": "0fee296a1b5542d778c99d64449222275aa79cdde262cd833deef00d33412e7d", "content_sha256": "09ca6d273aea07f3a9807e269f7620489f140ed07e40eaa73e16b5f385c61285", "fetched_at": "2026-08-17T00:01:59.181Z" }, { "role": "corroboration", "source_name": "METR — Measuring AI Ability to Complete Long Tasks", "source_url": "https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "source_date": "2025-03-19", "archive_url": "https://web.archive.org/web/2/https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "origin_org": "METR", "exact_quote_or_figure": "They outperform experts on most exam-style problems for a fraction of the cost.", "quote_sha256": "6a6fa336c63cdc31eb703e3b80816da7732b3ddc60230c2e308599c725774f1c", "content_sha256": "905599dd968d96c57eb63324b2ba0c176cd8cbb5f852ba8abba8b448c3253631", "fetched_at": "2026-08-17T00:01:58.392Z" }, { "role": "prediction", "source_name": "The Guardian / Observer (Kurzweil interview)", "source_url": "https://www.theguardian.com/technology/article/2024/jun/29/ray-kurzweil-google-ai-the-singularity-is-nearer", "source_date": "2024-06-29", "archive_url": "https://web.archive.org/web/2/https://www.theguardian.com/technology/article/2024/jun/29/ray-kurzweil-google-ai-the-singularity-is-nearer", "origin_org": "The Guardian / Observer (Kurzweil interview)", "exact_quote_or_figure": "I have stayed consistent. So 2029, both for human-level intelligence and for artificial general intelligence (AGI) – which is a little bit different.", "quote_sha256": "c9ca8818552383bf9ce200875d760c9199577cdb9f067c532286a2e74c28f5c9", "content_sha256": "f01e7678b74a56ffd63f48771da943038e963c2dd81966563dd809fb9f8c426e", "fetched_at": "2026-08-17T00:05:46.790Z" } ], "prediction_primary": null, "history": [ { "date": "2026-08-17", "change": "Target date 2029 is in the future, so the verdict is capped at on-track/partial. Two recognized-domain evidence items with different origin_orgs (>=1 measurement) document concrete positive progress…", "pr": "" }, { "date": "2026-11-03", "change": "The verdict 'on-track' remains correct (target 2029 is in the future). A fresh read of the Google DeepMind source confirms the concrete IMO 2025 measurement of progress toward human-level…", "pr": "" } ] }, { "id": "huang-ai-passes-every-test-by-2029", "url": "https://aiforecastledger.com/claims/huang-ai-passes-every-test-by-2029.html", "claim": "Within five years (by early 2029), AI does well on every single test the computer-science industry can put in front of it", "short_title": "AI passes every human test by 2029", "predictor": "Jensen Huang (Nvidia CEO)", "claimant_type": "person", "predicted_on": "2024-03", "target_date": "2029-03", "category": "benchmarks", "grading_rubric": "Met if by March 2029 frontier AI systems achieve strong performance on essentially every standardized human test the field proposes (bar exams, medical boards, olympiads, etc.); failed if significant test categories remain unconquered", "metric": "Coverage of standardized human tests at strong-pass level", "verdict": "on-track", "reviewer_status": "graded", "label": "On track", "last_reviewed": "2026-11-03", "current_measurement": { "metric_value": "35/42", "metric_unit": "IMO 2025 score (gold-medal standard)", "as_of": "2025-07-21", "source_quote": "solved five out of the six IMO problems perfectly, earning 35 total points", "quote_sha256": "0fee296a1b5542d778c99d64449222275aa79cdde262cd833deef00d33412e7d", "source_url": "https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/", "archive_url": "https://web.archive.org/web/20261103000938/https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/" }, "model_verifications": 2, "evidence": [ { "role": "measurement", "source_name": "Google DeepMind — Gemini Deep Think at IMO 2025", "source_url": "https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/", "source_date": "2025-07-21", "archive_url": "https://web.archive.org/web/20260816000453/https://deepmind.google/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/", "origin_org": "Google DeepMind", "exact_quote_or_figure": "solved five out of the six IMO problems perfectly, earning 35 total points", "quote_sha256": "0fee296a1b5542d778c99d64449222275aa79cdde262cd833deef00d33412e7d", "content_sha256": "09ca6d273aea07f3a9807e269f7620489f140ed07e40eaa73e16b5f385c61285", "fetched_at": "2026-08-16T00:01:33.626Z" }, { "role": "corroboration", "source_name": "METR — Measuring AI Ability to Complete Long Tasks", "source_url": "https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "source_date": "2025-03-19", "archive_url": "https://web.archive.org/web/20260816000451/https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "origin_org": "METR", "exact_quote_or_figure": "They outperform experts on most exam-style problems for a fraction of the cost.", "quote_sha256": "6a6fa336c63cdc31eb703e3b80816da7732b3ddc60230c2e308599c725774f1c", "content_sha256": "905599dd968d96c57eb63324b2ba0c176cd8cbb5f852ba8abba8b448c3253631", "fetched_at": "2026-08-16T00:01:32.844Z" }, { "role": "prediction", "source_name": "Fox Business (Stanford SIEPR Economic Summit remarks)", "source_url": "https://www.foxbusiness.com/technology/nvidia-ceo-jensen-huang-says-ai-could-pass-most-human-tests-5-years", "source_date": "2024-03-03", "archive_url": "https://web.archive.org/web/20260816000633/https://www.foxbusiness.com/technology/nvidia-ceo-jensen-huang-says-ai-could-pass-most-human-tests-5-years", "origin_org": "Fox Business (Stanford SIEPR Economic Summit remarks)", "exact_quote_or_figure": "\"If I gave an AI… every single test that you can possibly imagine, you make that list of tests and put it in front of the computer science industry, and I'm guessing in five years time, we'll do well on every single one,\" Huang said.", "quote_sha256": "44b92e99ac3850a434c013bd42e44c6b74ab716c1dcef1e5986d87ff5e62fb6c", "content_sha256": "5926269f98131fb97b88442e46c9b59823c9a3c448ef88551af46402e7a97b2a", "fetched_at": "2026-08-16T00:05:30.118Z" } ], "prediction_primary": { "kind": "recording", "url": "https://www.youtube.com/watch?v=cEg8cOx7UZk", "publisher": "Stanford Institute for Economic Policy Research (SIEPR)", "title": "Keynote by NVIDIA CEO Jensen Huang at 2024 SIEPR Economic Summit", "date": "2024-03", "archive_url": "https://web.archive.org/web/20260812183332/https://www.youtube.com/watch?v=cEg8cOx7UZk" }, "history": [ { "date": "2026-08-16", "change": "Target date 2029-03 is in the future, so on-track/partial is the ceiling. Google DeepMind (deepmind.google, measurement) records an officially graded, gold-medal-standard 35/42 on IMO 2025 — one of…", "pr": "" }, { "date": "2026-11-03", "change": "The verdict 'on-track' remains correct (target 2029-03 is in the future). A fresh read of the Google DeepMind source confirms the concrete IMO 2025 measurement of progress toward passing every test.", "pr": "" }, { "date": "2026-11-05", "change": "Linked the primary source: the organiser’s recording of the talk (Stanford SIEPR). The quote text still comes from the press report in the receipts. Verdict unchanged.", "pr": "" } ] }, { "id": "hassabis-agi-5-to-10-years-2025", "url": "https://aiforecastledger.com/claims/hassabis-agi-5-to-10-years-2025.html", "claim": "AGI arrives in the next five to ten years (2030–2035)", "short_title": "Hassabis: AGI in 5–10 years", "predictor": "Demis Hassabis (Google DeepMind)", "claimant_type": "person", "predicted_on": "2025-04", "target_date": "2035", "category": "timelines", "grading_rubric": "Met if a system widely accepted as AGI (general human-level capability across domains) is publicly demonstrated between 2030 and 2035; graded early as wrong only if Hassabis's stated criteria are clearly unmet by end of 2035", "metric": "Public demonstration of a generally human-level system", "verdict": "on-track", "reviewer_status": "graded", "label": "On track", "last_reviewed": "2026-11-03", "current_measurement": { "metric_value": "79.20", "metric_unit": "% resolved on SWE-bench Verified (top agent, bash-only default view)", "as_of": "2026-11-03", "source_quote": "Claude 4.5 Opus | Sonar Foundation Agent | 79.20", "quote_sha256": "aab3aa48de59ea386155a200e1848822a3d6ff65b292fc64b1da2f1dca989be3", "source_url": "https://www.swebench.com/", "archive_url": "https://web.archive.org/web/20261103000425/https://www.swebench.com/" }, "model_verifications": 2, "evidence": [ { "role": "measurement", "source_name": "SWE-bench Verified leaderboard", "source_url": "https://www.swebench.com/", "source_date": "2026-09-03", "archive_url": "https://web.archive.org/web/20260903000238/https://www.swebench.com/", "origin_org": "SWE-bench", "exact_quote_or_figure": "Claude 4.5 Opus | Sonar Foundation Agent | 79.20", "quote_sha256": "aab3aa48de59ea386155a200e1848822a3d6ff65b292fc64b1da2f1dca989be3", "content_sha256": "d0bfa531225a795cba3772e786c97f818ba55704b0894ef189a0d0b28731128c", "fetched_at": "2026-09-03T00:01:35.346Z" }, { "role": "measurement", "source_name": "METR — Measuring AI Ability to Complete Long Tasks", "source_url": "https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "source_date": "2025-03-19", "archive_url": "https://web.archive.org/web/20260903000436/https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "origin_org": "METR", "exact_quote_or_figure": "has been doubling approximately every 7 months for the last 6 years", "quote_sha256": "ca2a1c1ec39c6763103c94ae2b9893320aeabf670b60d40f648a64b2b2e023f0", "content_sha256": "905599dd968d96c57eb63324b2ba0c176cd8cbb5f852ba8abba8b448c3253631", "fetched_at": "2026-09-03T00:01:36.141Z" }, { "role": "prediction", "source_name": "CBS News — 60 Minutes interview transcript", "source_url": "https://www.cbsnews.com/news/artificial-intelligence-google-deepmind-ceo-demis-hassabis-60-minutes-transcript/", "source_date": "2025-04-20", "archive_url": "https://web.archive.org/web/20260903000507/https://www.cbsnews.com/news/artificial-intelligence-google-deepmind-ceo-demis-hassabis-60-minutes-transcript/", "origin_org": "CBS News — 60 Minutes interview transcript", "exact_quote_or_figure": "In the next five to ten years, I think.", "quote_sha256": "332223d541dfe7c7b0a494e28b000cfe07dd9478494182981b95a916f7c17ccd", "content_sha256": "9f7be9b3fc03f76ddac216319276e5f3e4f751c0ac350125e4036137b78f84a8", "fetched_at": "2026-09-03T00:05:00.809Z" } ], "prediction_primary": null, "history": [ { "date": "2026-09-03", "change": "Target date 2035 is well in the future, so on-track/partial is the ceiling. Two recognized-domain evidence items with different origin_orgs (both role:measurement) show concrete capability progress…", "pr": "" }, { "date": "2026-11-03", "change": "The verdict 'on-track' remains correct (target 2035 is in the future). A fresh read of the SWE-bench Verified leaderboard (fetched 2026-11-03) confirms the top bash-only entry remains at 79.20%…", "pr": "" } ] }, { "id": "lecun-no-human-level-ai-within-years-2024", "url": "https://aiforecastledger.com/claims/lecun-no-human-level-ai-within-years-2024.html", "claim": "Human-level AI is years away, if not a decade — and will not come from LLMs alone (counterpoint claim)", "short_title": "LeCun: human-level AI a decade out", "predictor": "Yann LeCun (then Meta Chief AI Scientist)", "claimant_type": "person", "predicted_on": "2024-10", "target_date": "2034", "category": "timelines", "grading_rubric": "Met if human-level AI does NOT appear before ~2030 (vindicating the skeptic call); failed if a widely accepted human-level system arrives well before the end of the decade or emerges primarily from LLM scaling", "metric": "Arrival date and architecture of the first accepted human-level system", "verdict": "on-track", "reviewer_status": "graded", "label": "On track", "last_reviewed": "2026-11-03", "current_measurement": { "metric_value": "79.20", "metric_unit": "% resolved on SWE-bench Verified (top agent, bash-only default view)", "as_of": "2026-11-03", "source_quote": "Claude 4.5 Opus | Sonar Foundation Agent | 79.20", "quote_sha256": "aab3aa48de59ea386155a200e1848822a3d6ff65b292fc64b1da2f1dca989be3", "source_url": "https://www.swebench.com/", "archive_url": "https://web.archive.org/web/20261103000425/https://www.swebench.com/" }, "model_verifications": 2, "evidence": [ { "role": "measurement", "source_name": "METR — Measuring AI Ability to Complete Long Tasks", "source_url": "https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "source_date": "2025-03-19", "archive_url": "https://web.archive.org/web/20260910000650/https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/", "origin_org": "METR", "exact_quote_or_figure": "the best AI agents are not currently able to carry out substantive projects by themselves or directly substitute for human labor.", "quote_sha256": "41ad689ea51059349b8ca8e34f36a9562063f865a8eafecc659555a1e649361d", "content_sha256": "905599dd968d96c57eb63324b2ba0c176cd8cbb5f852ba8abba8b448c3253631", "fetched_at": "2026-09-10T00:02:01.227Z" }, { "role": "measurement", "source_name": "SWE-bench Verified leaderboard", "source_url": "https://www.swebench.com/", "source_date": "unknown", "archive_url": "https://web.archive.org/web/20260910000433/https://www.swebench.com/", "origin_org": "SWE-bench", "exact_quote_or_figure": "Claude 4.5 Opus | Sonar Foundation Agent | 79.20", "quote_sha256": "aab3aa48de59ea386155a200e1848822a3d6ff65b292fc64b1da2f1dca989be3", "content_sha256": "d0bfa531225a795cba3772e786c97f818ba55704b0894ef189a0d0b28731128c", "fetched_at": "2026-09-10T00:02:00.375Z" }, { "role": "prediction", "source_name": "TechCrunch (Hudson Forum keynote)", "source_url": "https://techcrunch.com/2024/10/16/metas-ai-chief-says-world-models-are-key-to-human-level-ai-but-it-might-be-10-years-out/", "source_date": "2024-10-16", "archive_url": "https://web.archive.org/web/20260910000757/https://techcrunch.com/2024/10/16/metas-ai-chief-says-world-models-are-key-to-human-level-ai-but-it-might-be-10-years-out/", "origin_org": "TechCrunch (Hudson Forum keynote)", "exact_quote_or_figure": "“It’s going to take years before we can get everything here to work, if not a decade,” said LeCun. “Mark Zuckerberg keeps asking me how long it’s going to take.”", "quote_sha256": "7de83edc8fd3ba8728cb9bd630f849afd267b5e1ba1a2ea3d96567233eea395a", "content_sha256": "36cae2bf528ad8cfd137c86eccd551b08781535620ec53109af922183b033a58", "fetched_at": "2026-09-10T00:07:06.319Z" } ], "prediction_primary": { "kind": "recording", "url": "https://www.youtube.com/watch?v=4DsCtgtQlZU", "publisher": "Hudson Forum", "title": "Keynote: Yann LeCun, “Human-Level AI”", "date": "2024-10", "archive_url": "https://web.archive.org/web/20260527063735/https://www.youtube.com/watch?v=4DsCtgtQlZU" }, "history": [ { "date": "2026-09-10", "change": "Target date 2034 is far in the future, so on-track/partial is the ceiling. LeCun's skeptic call is currently holding: METR (metr.org, measurement, different org) measures that the best agents still…", "pr": "" }, { "date": "2026-11-03", "change": "The verdict 'on-track' remains correct (target 2034 is in the future). A fresh read of the SWE-bench Verified leaderboard (fetched 2026-11-03) confirms the top bash-only entry remains at 79.20%…", "pr": "" }, { "date": "2026-11-05", "change": "Linked the primary source: the organiser’s recording of the talk (Hudson Forum). The quote text still comes from the press report in the receipts. Verdict unchanged.", "pr": "" } ] } ] }