https://scorecard.aiforecastledger.com We log the loudest public predictions about AI, then grade each one against dated, linkable evidence. No vibes: only receipts. Tue, 28 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com/notes/note-checkpoint-superhuman-coder-2027-2026-01.html https://scorecard.aiforecastledger.com/notes/note-checkpoint-superhuman-coder-2027-2026-01.html The AI-2027 report predicted a superhuman coder by end of 2027 (predicted 2025-04, target 2027-12). The early-2026 checkpoint tracks SWE-bench Verified performance, which the leaderboard shows at 76.80% for Claude 4.5 Opus (high reasoning) as of July 15, 2026. The final 2027-12 milestone remains pending. METR's March 2025 long-task assessment noted that AI agents could not yet carry out substantive projects independently. Tue, 28 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com/notes/note-anniversary-hinton-stop-training-radiologists-by-2021-2026.html https://scorecard.aiforecastledger.com/notes/note-anniversary-hinton-stop-training-radiologists-by-2021-2026.html Geoffrey Hinton predicted in 2016 that deep learning would outperform radiologists within five years (target 2021) and that training of radiologists should stop. The Harvey L. Neiman Health Policy Institute reports the US radiologist workforce grew 17.3% (30,723 to 36,024) from 2014 to 2023, and a 2025 Nature paper documents a persistent radiology workforce shortage. The target date of 2021 is now five years past. Mon, 27 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com/notes/note-checkpoint-amodei-ai-writes-90pct-of-code-by-2025-2025-09.html https://scorecard.aiforecastledger.com/notes/note-checkpoint-amodei-ai-writes-90pct-of-code-by-2025-2025-09.html Dario Amodei stated in March 2025 that AI could write 90% of code within three to six months. By the September 2025 checkpoint, an arXiv study estimated AI wrote 30.1% of Python functions from US contributors as of December 2024, and Microsoft reported 20–30% of its code was AI-generated. No verified industry-wide 90% figure surfaced by that checkpoint, per a Skeptics StackExchange survey. Mon, 27 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#superhuman-coder-2027 superhuman-coder-2027 On track — Met when a model autonomously completes a non-trivial PR end-to-end at senior-eng level Wed, 15 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#altman-agents-join-workforce-2025 altman-agents-join-workforce-2025 Missed — 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 Wed, 15 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#musk-ai-smarter-than-any-human-2025 musk-ai-smarter-than-any-human-2025 Missed — 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 Sun, 05 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#amodei-ai-writes-90pct-of-code-by-2025 amodei-ai-writes-90pct-of-code-by-2025 Missed — 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 Fri, 03 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#hinton-stop-training-radiologists-by-2021 hinton-stop-training-radiologists-by-2021 Missed — Met if AI had displaced human radiologists (falling demand / headcount) by ~2021; failed if the workforce kept growing and AI became a complement Thu, 02 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#kurzweil-turing-test-by-2029 kurzweil-turing-test-by-2029 Pending — 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 Thu, 02 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#huang-ai-passes-every-test-by-2029 huang-ai-passes-every-test-by-2029 Pending — 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 Thu, 02 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#hassabis-agi-5-to-10-years-2025 hassabis-agi-5-to-10-years-2025 Pending — 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 Thu, 02 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#lecun-no-human-level-ai-within-years-2024 lecun-no-human-level-ai-within-years-2024 Pending — 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 Thu, 02 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com/notes/note-anniversary-ai-imo-gold-medal-by-2025-2026.html https://scorecard.aiforecastledger.com/notes/note-anniversary-ai-imo-gold-medal-by-2025-2026.html The claim by Eliezer Yudkowsky, made in a 2022 bet with Paul Christiano, targeted an AI winning an IMO gold medal by 2025. The target date has passed. Google DeepMind announced in July 2025 that an advanced version of Gemini Deep Think scored 35/42 at the 2025 IMO, meeting the gold-medal cutoff. Thu, 02 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com/notes/note-anniversary-metaculus-weakly-general-ai-2028-2026.html https://scorecard.aiforecastledger.com/notes/note-anniversary-metaculus-weakly-general-ai-2028-2026.html The Metaculus community forecast, made in 2020, predicts the public announcement of the first weakly general AI system around 2028. The target date is still two years away. The ledger has not yet taken a verified read of the live community median; the claim remains undetermined pending one. Thu, 02 Jul 2026 00:00:00 GMT https://scorecard.aiforecastledger.com/notes/note-anniversary-superhuman-coder-2027-2026.html https://scorecard.aiforecastledger.com/notes/note-anniversary-superhuman-coder-2027-2026.html The 'superhuman coder by end of 2027' prediction from the AI-2027 report was made in April 2025. As of this anniversary, the target date of December 2027 is still 18 months away. The prediction's associated checkpoints show progress from 'stumbling agents' in mid-2025 to coding agents exceeding 85% on SWE-bench Verified by mid-2026. Sun, 28 Jun 2026 00:00:00 GMT https://scorecard.aiforecastledger.com/notes/note-checkpoint-superhuman-coder-2027-2025-06.html https://scorecard.aiforecastledger.com/notes/note-checkpoint-superhuman-coder-2027-2025-06.html The AI-2027 scenario forecast predicted the appearance of first usable AI coding agents by mid-2025. This checkpoint was assessed as 'hit' in June 2026, confirming that coding agents emerged and began autonomously resolving real GitHub issues on the SWE-bench Verified benchmark. The source for this assessment is the SWE-bench leaderboard. Sun, 28 Jun 2026 00:00:00 GMT https://scorecard.aiforecastledger.com/notes/note-anniversary-amodei-ai-writes-90pct-of-code-by-2025-2026.html https://scorecard.aiforecastledger.com/notes/note-anniversary-amodei-ai-writes-90pct-of-code-by-2025-2026.html In March 2025 Dario Amodei said AI would be writing 90% of code within three to six months. That window closed in March 2026. The claim is past deadline and awaiting a formal grade. Sat, 27 Jun 2026 00:00:00 GMT https://scorecard.aiforecastledger.com/notes/note-anniversary-hinton-stop-training-radiologists-by-2021-2021.html https://scorecard.aiforecastledger.com/notes/note-anniversary-hinton-stop-training-radiologists-by-2021-2021.html In 2016 Geoffrey Hinton said we should stop training radiologists because deep learning would outperform them within five years. The 2021 deadline passed; radiologist demand has since grown rather than collapsed. The claim is past its window and awaiting a formal grade — not yet a verdict. Sat, 27 Jun 2026 00:00:00 GMT https://scorecard.aiforecastledger.com/notes/note-methodology-receipts-2026-06.html https://scorecard.aiforecastledger.com/notes/note-methodology-receipts-2026-06.html Every resolved claim on the ledger is graded against dated, linkable evidence, blind-verified by two independent models, and bound to a commit in version history. We publish the receipt, not the take. Sat, 27 Jun 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#ai-imo-gold-medal-by-2025 ai-imo-gold-medal-by-2025 Hit — Met when, at an IMO through 2025, an AI built beforehand scores at or above the gold-medal cutoff under competition conditions Thu, 25 Jun 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#metaculus-weakly-general-ai-2028 metaculus-weakly-general-ai-2028 Pending — Met if a system meeting the Metaculus weakly-general-AI resolution criteria is publicly announced near the community median Wed, 24 Jun 2026 00:00:00 GMT https://scorecard.aiforecastledger.com#agent-economy-2026 agent-economy-2026 Pending — Met when a public company attributes >1% revenue to autonomous agents Sat, 20 Jun 2026 00:00:00 GMT