{ "name": "Singularity Tracker", "publisher": "Prime Signal", "url": "https://primesignal.org/singularity-tracker.html", "license": "CC BY 4.0", "cite": "Prime Signal, Singularity Tracker, updated 29 Oct 2026", "summary": { "updated": "2026-10-29", "edition": "v1.1", "previous_update": "2026-08-20", "index": 52.0, "verified_index": 29.0, "previous_index": 51.0, "previous_verified_index": 28.0, "milestones_achieved": 50, "milestones_partial": 4, "by_verification": { "replicated": 13, "peer_reviewed": 17, "preprint": 9, "claimed": 15 }, "gatekeepers_signed_off": 1, "gatekeepers": 3, "superintelligence": 57.3, "verified_superintelligence": 41.7 }, "fields": [ { "id": "math", "name": "Mathematics", "short": "Math", "stage": 3, "latency_log10_hours": -3.5, "latency": "Seconds", "latency_detail": "Seconds with a proof checker (Lean); months for peer review", "tracks": "Competition math, machine-checked proofs, open problems, journal acceptance.", "bottom_line": "Journals now publish AI-made proofs. The biggest 2026 claims, including Navier–Stokes, are not refereed yet.", "held_back": "Peer review can't keep pace with how fast AI produces results.", "score": 85, "verified_score": 70, "previous_score": 75, "previous_verified_score": 60, "milestones": [ { "field_id": "math", "n": 1, "milestone": "AI matches the best humans at competition math", "status": "achieved", "verification": "replicated", "reached": "2025-07", "outlook": "", "evidence": "IMO gold under official grading, then perfect scores in 2026", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "math", "n": 2, "milestone": "AI writes machine-checked proofs of known results at research scale", "status": "achieved", "verification": "replicated", "reached": "2025-09", "outlook": "", "evidence": "Strong Prime Number Theorem formalized in Lean in three weeks", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "math", "n": 3, "milestone": "AI solves a genuinely open problem, with a machine-checked proof", "status": "achieved", "verification": "replicated", "reached": "2026-01-07", "outlook": "", "evidence": "Erdős problem #728, Lean proof", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "erdos728" ], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "math", "n": 4, "milestone": "AI disproves a famous conjecture, confirmed by senior mathematicians", "status": "achieved", "verification": "replicated", "reached": "2026-05-20", "outlook": "", "evidence": "Unit distance conjecture (1946), checked by nine mathematicians", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "unit", "unit_oai" ], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "math", "n": 5, "milestone": "A second lab's AI disproves a major conjecture", "status": "achieved", "verification": "replicated", "reached": "2026-07-21", "outlook": "", "evidence": "Jacobian conjecture false in dimension 3. Reviewed by Tao, rechecked in Isabelle", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "jacobian", "jacobian_afp" ], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "math", "n": 6, "milestone": "A batch of open problems solved at once, with public machine-checked proofs", "status": "achieved", "verification": "preprint", "reached": "2026-08-01", "outlook": "", "evidence": "OpenAI: ten results in Lean (Aug), then 721 manuscripts, about 42% with Lean proofs (Oct). No outside audit yet", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "astra_gh", "oai_repo" ], "states": [ [ "2026-08-20", "achieved", "preprint" ] ] }, { "field_id": "math", "n": 7, "milestone": "AI makes a major advance on a first-rank problem", "status": "achieved", "verification": "replicated", "reached": "2026-08-10", "outlook": "", "evidence": "Zeta zeros proven on the critical line: 41.6% to 67.2%. A simpler human proof followed", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "riemann", "riemann_paper", "lamzouri" ], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "math", "n": 8, "milestone": "A journal publishes an AI-led result after peer review", "status": "achieved", "verification": "peer_reviewed", "reached": "2026-05", "outlook": "", "evidence": "First AI-proved papers in journals (Axiom Math). Science published AlphaProof Nexus on 8 Oct 2026", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 10, "verified_points": 10, "sources": [ "nexus", "axiom1", "axiom5" ], "states": [ [ "2026-08-20", "partial", "claimed" ], [ "2026-10-29", "achieved", "peer_reviewed" ] ] }, { "field_id": "math", "n": 9, "milestone": "AI develops genuinely new theory", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2031", "evidence": "Watch how experts judge OpenAI's local-smoothing proof", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "oai_oct26" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "math", "n": 10, "milestone": "AI is the main prover on a Millennium Prize Problem", "status": "partial", "verification": "preprint", "reached": "2026-09-08", "outlook": "", "evidence": "Navier–Stokes, with a Lean proof. Not refereed; no Clay ruling", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 5, "verified_points": 0, "sources": [ "navier" ], "states": [ [ "2026-10-29", "partial", "preprint" ] ] } ] }, { "id": "ai", "name": "AI / Recursion", "short": "AI", "stage": 3, "latency_log10_hours": 0.0, "latency": "Hours", "latency_detail": "Seconds to days: tests, benchmarks, production trials", "tracks": "Coding agents, AI-found algorithms, AI improving the chips and code it runs on.", "bottom_line": "AI now speeds up the work of building AI. It doesn't yet invent new AI on its own.", "held_back": "Research taste: AI ideas that score well in the lab rarely hold up in production.", "score": 60, "verified_score": 10, "previous_score": 60, "previous_verified_score": 10, "milestones": [ { "field_id": "ai", "n": 1, "milestone": "AI agents solve over half of real coding tasks (SWE-bench Verified)", "status": "achieved", "verification": "claimed", "reached": "2024-12", "outlook": "", "evidence": "o3 reported 71.7% (Dec 2024). Latest: 81.4%, Claude Opus 4.6 (Feb 2026). Lab-reported", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "ai", "n": 2, "milestone": "An AI-discovered learning algorithm beats the best human one, in a journal", "status": "achieved", "verification": "peer_reviewed", "reached": "2025-10-22", "outlook": "", "evidence": "DiscoRL", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "discorl" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "ai", "n": 3, "milestone": "AI-found optimizations run in production at a frontier lab", "status": "achieved", "verification": "claimed", "reached": "2025-05-14", "outlook": "", "evidence": "AlphaEvolve recovered 0.7% of Google's compute", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "alphaevolve" ], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "ai", "n": 4, "milestone": "AI improves the chips or code it runs on", "status": "achieved", "verification": "claimed", "reached": "2025-05-14", "outlook": "", "evidence": "A circuit change in a TPU; 32.5% faster attention kernel", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "alphaevolve" ], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "ai", "n": 5, "milestone": "A model cuts the cost of serving itself", "status": "achieved", "verification": "claimed", "reached": "2026-07-29", "outlook": "", "evidence": "GPT-5.6 Sol: 20% lower serving cost, done with Codex", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "sol" ], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "ai", "n": 6, "milestone": "AI writes most of the code at a frontier lab (over 75%)", "status": "achieved", "verification": "claimed", "reached": "2026-05", "outlook": "", "evidence": "Over 80% at Anthropic", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "rsi" ], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "ai", "n": 7, "milestone": "An AI research result holds up in production", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2030", "evidence": "A 0.97 lab score shrank to noise in production", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "w2s" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "ai", "n": 8, "milestone": "A frontier lab shows AI has at least doubled its research speed", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2031", "evidence": "Anthropic said no as of Apr 2026", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "ai", "n": 9, "milestone": "A frontier model designed end to end by AI", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2030 to 2035", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "ai", "n": 10, "milestone": "AI designs and builds its own successor", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2032 to 2040+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] } ] }, { "id": "physics", "name": "Physics", "short": "Physics", "stage": 2, "latency_log10_hours": 5.24, "latency": "Decades", "latency_detail": "Years to decades: build a new detector or collider", "tracks": "AI in instruments, rediscovered and corrected laws, predictions confirmed by experiment.", "bottom_line": "AI finds real results where physics can be simulated. New fundamental physics needs new experiments, decades away.", "held_back": "Experiments take years to decades to build.", "score": 55, "verified_score": 30, "previous_score": 55, "previous_verified_score": 30, "milestones": [ { "field_id": "physics", "n": 1, "milestone": "AI rediscovers known physical laws from raw data", "status": "achieved", "verification": "preprint", "reached": "2025-04", "outlook": "", "evidence": "Newton's second law, energy conservation and more, through Mar 2026", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "achieved", "preprint" ] ] }, { "field_id": "physics", "n": 2, "milestone": "AI runs in the control loop of a frontier instrument", "status": "achieved", "verification": "peer_reviewed", "reached": "2025-09-04", "outlook": "", "evidence": "LIGO: 30 to 100x less control noise", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "ligo" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "physics", "n": 3, "milestone": "An AI-designed instrument beats the human design", "status": "achieved", "verification": "peer_reviewed", "reached": "2025-04-14", "outlook": "", "evidence": "Urania detector designs", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "urania" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "physics", "n": 4, "milestone": "AI corrects an accepted physical model, confirmed by experiment", "status": "achieved", "verification": "replicated", "reached": "2025-07", "outlook": "", "evidence": "Dusty plasma: two standing assumptions overturned", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "plasma" ], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "physics", "n": 5, "milestone": "AI finds new mathematical objects in physics equations", "status": "achieved", "verification": "preprint", "reached": "2025-09-18", "outlook": "", "evidence": "New fluid singularities (Sep 2025). Plasma counterexamples, preprint (Sep 2026)", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "euler", "grad", "grad_umd" ], "states": [ [ "2026-08-20", "achieved", "preprint" ] ] }, { "field_id": "physics", "n": 6, "milestone": "AI matches expert physicists on a research-grade test", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2027 to 2029", "evidence": "CritPt 32.3%, up from 5.7% in 11 months", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "critpt" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "physics", "n": 7, "milestone": "AI does publishable research end to end, above grad-student level", "status": "partial", "verification": "preprint", "reached": "2026-01", "outlook": "", "evidence": "A paper its own author rated second-year-grad level", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 5, "verified_points": 0, "sources": [ "scet" ], "states": [ [ "2026-08-20", "partial", "preprint" ] ] }, { "field_id": "physics", "n": 8, "milestone": "AI predicts new fundamental physics, later confirmed by experiment", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2030s", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "physics", "n": 9, "milestone": "AI produces a new fundamental law the field adopts", "status": "not_yet", "verification": "", "reached": "", "outlook": "No credible window", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "physics", "n": 10, "milestone": "AI contributes to a confirmed result beyond the Standard Model", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2040s", "evidence": "Waits on new colliders", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] } ] }, { "id": "chips", "name": "Chips & Compute", "short": "Chips", "stage": 2, "latency_log10_hours": 3.34, "latency": "Months", "latency_detail": "Simulation in days; three to six months to make the chip", "tracks": "AI in chip design tools, AI-designed circuit blocks, autonomous chip design.", "bottom_line": "AI improves parts of chips that ship today. Humans still design and sign off on the whole chip.", "held_back": "Memory and power supply, and human sign-off on verification.", "score": 45, "verified_score": 25, "previous_score": 45, "previous_verified_score": 25, "milestones": [ { "field_id": "chips", "n": 1, "milestone": "AI optimizers are standard in commercial chip design tools", "status": "achieved", "verification": "claimed", "reached": "2024-03", "outlook": "", "evidence": "Synopsys: hundreds of AI-enabled tapeouts (Mar 2024). Cadence: over 1,000 (Feb 2026). Unaudited", "recorded": "2026-08-20", "updated": "2026-10-29", "points": 10, "verified_points": 0, "sources": [ "synopsys", "cadence" ], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "chips", "n": 2, "milestone": "The cost of a fixed AI capability falls over 100x in under two years", "status": "achieved", "verification": "replicated", "reached": "2024-12", "outlook": "", "evidence": "$37.50 to $0.12 per million tokens, Mar 2023 to Dec 2024", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "epoch_price" ], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "chips", "n": 3, "milestone": "An AI-designed chip layout ships in production", "status": "partial", "verification": "peer_reviewed", "reached": "2021-06", "outlook": "", "evidence": "AlphaChip layouts in Google TPUs since 2020. Peer-reviewed but disputed", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 5, "verified_points": 5, "sources": [ "alphachip" ], "states": [ [ "2026-08-20", "partial", "peer_reviewed" ] ] }, { "field_id": "chips", "n": 4, "milestone": "An AI-written circuit change ships in a production accelerator", "status": "achieved", "verification": "claimed", "reached": "2025-05-14", "outlook": "", "evidence": "AlphaEvolve", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "alphaevolve" ], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "chips", "n": 5, "milestone": "An AI-designed radio-frequency block beats human designs when built", "status": "achieved", "verification": "peer_reviewed", "reached": "2024-12-30", "outlook": "", "evidence": "Designed in minutes rather than weeks", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "rfblock" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "chips", "n": 6, "milestone": "Chip verification signed off with no human review", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2029 to 2032", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "semieng" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "chips", "n": 7, "milestone": "A full production chip designed with no human intervention", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2030 to 2034", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "semieng" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "chips", "n": 8, "milestone": "An AI-originated chip architecture beats the human one it replaces", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2031 to 2035", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "chips", "n": 9, "milestone": "AI designs, verifies and builds its own next accelerator", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2035+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "chips", "n": 10, "milestone": "AI runs the compute supply chain end to end", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2035+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] } ] }, { "id": "bio", "name": "Health & Bio", "short": "Bio", "stage": 2, "latency_log10_hours": 4.72, "latency": "Years", "latency_detail": "3 to 10 years: lab work, three trial phases, then a regulator", "tracks": "Protein design, AI in clinical care, AI-designed drugs in trials and approvals.", "bottom_line": "AI-designed drugs are in trials, one in Phase III. None is approved yet.", "held_back": "Clinical trials take years, and AI can't shorten them much.", "score": 50, "verified_score": 40, "previous_score": 50, "previous_verified_score": 40, "milestones": [ { "field_id": "bio", "n": 1, "milestone": "AI protein structure prediction is standard research infrastructure", "status": "achieved", "verification": "replicated", "reached": "2021-07", "outlook": "", "evidence": "AlphaFold database opens. Independently benchmarked; 2024 Nobel Prize", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "bio", "n": 2, "milestone": "AI improves a clinical outcome in a randomized trial", "status": "achieved", "verification": "peer_reviewed", "reached": "2025", "outlook": "", "evidence": "MASAI, The Lancet, 100,000+ women: 9% more cancers found", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "bio", "n": 3, "milestone": "The first AI-originated drug enters Phase I", "status": "achieved", "verification": "claimed", "reached": "2020-01", "outlook": "", "evidence": "Exscientia DSP-1181", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "bio", "n": 4, "milestone": "AI designs new proteins with atomic accuracy", "status": "achieved", "verification": "peer_reviewed", "reached": "2025-11-05", "outlook": "", "evidence": "5 Nov 2025", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "protein" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "bio", "n": 5, "milestone": "A generative-AI drug reaches Phase III", "status": "achieved", "verification": "peer_reviewed", "reached": "2026-07-07", "outlook": "", "evidence": "Rentosertib, 320 patients; first patient dosed 10 Sep 2026", "recorded": "2026-08-20", "updated": "2026-10-29", "points": 10, "verified_points": 10, "sources": [ "rento", "rento_dose", "rento_p2" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "bio", "n": 6, "milestone": "A regulator approves a generative-AI-designed drug", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected around 2030", "evidence": "None on the FDA's 2026 list", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "fda_novel" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "bio", "n": 7, "milestone": "AI beats doctors on real patient outcomes, in a prospective trial", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2029 to 2033", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "bio", "n": 8, "milestone": "A virtual cell model reliably predicts how cells respond", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2031", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "bio", "n": 9, "milestone": "A disease durably cured by an AI-designed therapy", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2032 to 2035+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "bio", "n": 10, "milestone": "AI runs drug discovery, trial design and readout end to end", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2035+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] } ] }, { "id": "mat", "name": "Materials & Chemistry", "short": "Materials", "stage": 2, "latency_log10_hours": 2.86, "latency": "Weeks", "latency_detail": "Weeks to months to make and test; years to scale up", "tracks": "AI-predicted materials, lab synthesis, autonomous labs, materials in products.", "bottom_line": "AI predicts many new materials. Few have been made and confirmed by others, and some big claims were walked back.", "held_back": "It's hard to confirm what a lab actually made, and training data is messy.", "score": 50, "verified_score": 40, "previous_score": 50, "previous_verified_score": 40, "milestones": [ { "field_id": "mat", "n": 1, "milestone": "AI-curated materials databases are standard infrastructure", "status": "achieved", "verification": "replicated", "reached": "2023", "outlook": "", "evidence": "Materials Project: 650,000 users", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "mat", "n": 2, "milestone": "AI generates stable crystal candidates at scale", "status": "achieved", "verification": "peer_reviewed", "reached": "2023-11-29", "outlook": "", "evidence": "GNoME, 2.2M structures. Disputed: 83,000+ removed; duplication preprint Oct 2026", "recorded": "2026-08-20", "updated": "2026-10-29", "points": 10, "verified_points": 10, "sources": [ "gnome_dup" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "mat", "n": 3, "milestone": "An AI-predicted material is made and its target property measured", "status": "achieved", "verification": "peer_reviewed", "reached": "2025-01-16", "outlook": "", "evidence": "MatterGen: 158 GPa against a 200 GPa target", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "mattergen" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "mat", "n": 4, "milestone": "An outside group makes an AI-predicted compound and confirms it", "status": "achieved", "verification": "preprint", "reached": "2026-03-24", "outlook": "", "evidence": "MnFeCo4Si2", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "mnfeco" ], "states": [ [ "2026-08-20", "achieved", "preprint" ] ] }, { "field_id": "mat", "n": 5, "milestone": "AI predicts a new superconductor, then it's made and confirmed", "status": "achieved", "verification": "peer_reviewed", "reached": "2026-06-17", "outlook": "", "evidence": "YRu3B2 and LuRu3B2, confirmed four ways at Rice", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "kagome" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "mat", "n": 6, "milestone": "An autonomous lab runs a multi-week discovery campaign with no humans", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2029 to 2033", "evidence": "Humans still move samples", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "lila" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "mat", "n": 7, "milestone": "An AI-discovered material ships in a commercial product", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2032", "evidence": "Discovered Materials claims it; no data published", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "dmat" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "mat", "n": 8, "milestone": "An AI-designed catalyst is used in industry", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2029 to 2034", "evidence": "Best public result: 20 millilitres", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "catalyst" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "mat", "n": 9, "milestone": "Discovery to qualified material in under three years", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2033+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "mat", "n": 10, "milestone": "AI finds a room-temperature superconductor", "status": "not_yet", "verification": "", "reached": "", "outlook": "After 2035, maybe never", "evidence": "AI finds so far top out at 6.5 K", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "elements" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] } ] }, { "id": "robo", "name": "Robotics", "short": "Robotics", "stage": 2, "latency_log10_hours": 3.94, "latency": "A year", "latency_detail": "Trials run daily; proving reliability takes years", "tracks": "Robot AI models, humanoid cost, robotaxis, real-world deployments.", "bottom_line": "Robotaxis work at scale. Robots that handle objects are still unreliable: 46% success picking items off the floor.", "held_back": "Too little training data, and reliability far below what paid work needs.", "score": 50, "verified_score": 10, "previous_score": 50, "previous_verified_score": 10, "milestones": [ { "field_id": "robo", "n": 1, "milestone": "Robot AI models transfer to new robot bodies with little data", "status": "achieved", "verification": "claimed", "reached": "2026-07-30", "outlook": "", "evidence": "Gemini Robotics 2: under 200 examples", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "gr2" ], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "robo", "n": 2, "milestone": "Full-size humanoids cost under $20,000", "status": "achieved", "verification": "replicated", "reached": "2025-07", "outlook": "", "evidence": "Unitree R1 from $4,900. Public prices", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "robo", "n": 3, "milestone": "A robotaxi service gives 250,000+ paid driverless rides a week", "status": "achieved", "verification": "claimed", "reached": "2026-02", "outlook": "", "evidence": "Waymo: 400,000+ a week. Company figure", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "robo", "n": 4, "milestone": "A driverless fleet proves safer than humans over 100M miles", "status": "achieved", "verification": "claimed", "reached": "2026-03", "outlook": "", "evidence": "Waymo: 220.6M miles, 94% fewer serious-injury crashes. Company data", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "waymo" ], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "robo", "n": 5, "milestone": "A robot runs a full commercial shift over 95% autonomously", "status": "achieved", "verification": "claimed", "reached": "2026-02", "outlook": "", "evidence": "96.4% autonomy", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "robo", "n": 6, "milestone": "Humanoids work real shifts at customer sites, at scale, without remote control", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2031", "evidence": "Largest so far: 3 to 10 units", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "robo", "n": 7, "milestone": "10,000 humanoids in commercial use", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2030", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "tesla" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "robo", "n": 8, "milestone": "Robots handle new objects and tasks with over 80% success", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2030", "evidence": "Today: floor pick 45.7%, dustpan 32%", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "gr2" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "robo", "n": 9, "milestone": "Physical labor automated at economic scale", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2032 to 2035+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "robo", "n": 10, "milestone": "A robot performs a full surgery on its own", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2030+", "evidence": "All 3.1M da Vinci procedures in 2025 were human-run", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] } ] }, { "id": "econ", "name": "Economy & Labor", "short": "Economy", "stage": 2, "latency_log10_hours": 4.76, "latency": "Years", "latency_detail": "5 to 8 years: adoption spreads, then shows up in national data", "tracks": "Business adoption, AI revenue, job effects, productivity, AI agent payments.", "bottom_line": "About 1 in 4 US firms use AI. It doesn't show up yet in productivity or jobs data.", "held_back": "New technology takes years to spread and show up in national statistics.", "score": 50, "verified_score": 20, "previous_score": 50, "previous_verified_score": 20, "milestones": [ { "field_id": "econ", "n": 1, "milestone": "At least 20% of US firms use AI", "status": "achieved", "verification": "replicated", "reached": "2026-07-26", "outlook": "", "evidence": "Census survey: 21.8% (two weeks to 26 Jul 2026). Latest: 24.6% (two weeks to 4 Oct)", "recorded": "2026-08-20", "updated": "2026-10-29", "points": 10, "verified_points": 10, "sources": [ "btos" ], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "econ", "n": 2, "milestone": "A frontier AI lab passes $50B annual revenue", "status": "achieved", "verification": "claimed", "reached": "2026-07", "outlook": "", "evidence": "Anthropic past $65B. Company figure", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "econ", "n": 3, "milestone": "AI matches experts on half of real work tasks", "status": "achieved", "verification": "preprint", "reached": "2025-09-25", "outlook": "", "evidence": "GDPval, 44 occupations, blind grading", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "achieved", "preprint" ] ] }, { "field_id": "econ", "n": 4, "milestone": "A job effect shows up in payroll data", "status": "achieved", "verification": "preprint", "reached": "2026-06", "outlook": "", "evidence": "Workers aged 22 to 25 in AI-exposed jobs 19% below trend. Working paper", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "canaries" ], "states": [ [ "2026-08-20", "achieved", "preprint" ] ] }, { "field_id": "econ", "n": 5, "milestone": "AI becomes the top stated reason for announced job cuts", "status": "achieved", "verification": "replicated", "reached": "2026-03", "outlook": "", "evidence": "Five straight months, Mar to Jul 2026, though small against 158.9M jobs", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [], "states": [ [ "2026-08-20", "achieved", "replicated" ] ] }, { "field_id": "econ", "n": 6, "milestone": "Productivity grows above trend for three years", "status": "not_yet", "verification": "", "reached": "", "outlook": "Going the wrong way", "evidence": "0.8% in 2025", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "bls" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "econ", "n": 7, "milestone": "Job losses show up broadly in national data", "status": "not_yet", "verification": "", "reached": "", "outlook": "Not seen through Q1 2026", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "econ", "n": 8, "milestone": "AI agent spending reaches material scale", "status": "not_yet", "verification": "", "reached": "", "outlook": "Measurable, not material", "evidence": "x402: about $54M since launch, self-reported", "recorded": "2026-08-20", "updated": "2026-10-29", "points": 0, "verified_points": 0, "sources": [ "x402", "chainalysis" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "econ", "n": 9, "milestone": "A developed economy grows 4%+ in a year", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2031+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "econ", "n": 10, "milestone": "Labor's share of income falls measurably", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2031+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] } ] }, { "id": "neuro", "name": "Neuro & Interface", "short": "Neuro", "stage": 2, "latency_log10_hours": 4.24, "latency": "Years", "latency_detail": "1 to 3 years: surgery plus a long clinical study", "tracks": "Brain-computer interfaces, speech decoding, approvals, patient counts.", "bottom_line": "Brain implants let paralyzed people talk and type at home. All are still experimental.", "held_back": "Low bandwidth, and no way to write information into the brain.", "score": 45, "verified_score": 35, "previous_score": 45, "previous_verified_score": 35, "milestones": [ { "field_id": "neuro", "n": 1, "milestone": "AI makes high-bandwidth brain decoding practical", "status": "achieved", "verification": "peer_reviewed", "reached": "2023-08", "outlook": "", "evidence": "Speech decoding from implants in Nature. Now 56 words a minute sustained", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "bci" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "neuro", "n": 2, "milestone": "Brain-implant cursor control matches able-bodied speed", "status": "achieved", "verification": "claimed", "reached": "2026-01", "outlook": "", "evidence": "Over 10 bits per second", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "neuro", "n": 3, "milestone": "A brain implant is used at home, unsupervised, for years", "status": "achieved", "verification": "peer_reviewed", "reached": "2026-06-15", "outlook": "", "evidence": "3,800+ hours. One participant", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "bci" ], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "neuro", "n": 4, "milestone": "AI decodes imagined speech", "status": "achieved", "verification": "peer_reviewed", "reached": "2025-08-14", "outlook": "", "evidence": "Cell, four participants", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "neuro", "n": 5, "milestone": "AI-driven brain mapping yields a new neuroscience finding", "status": "partial", "verification": "peer_reviewed", "reached": "2026-06-08", "outlook": "", "evidence": "Fly connectome: AI aligned the data, humans did the science", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 5, "verified_points": 5, "sources": [], "states": [ [ "2026-08-20", "partial", "peer_reviewed" ] ] }, { "field_id": "neuro", "n": 6, "milestone": "A brain implant restores conversation-speed speech (140+ words a minute)", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2031", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "neuro", "n": 7, "milestone": "FDA approves a permanent high-channel brain implant", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected around 2030", "evidence": "Paradromics trial expanded Oct 2026", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "paradromics" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "neuro", "n": 8, "milestone": "1,000 patients implanted", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2029 to 2032", "evidence": "About 150 today", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "neuro", "n": 9, "milestone": "Imagined speech decoded without surgery", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2032+, maybe never", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "neuro", "n": 10, "milestone": "Writing information, not just sensation, into a brain", "status": "not_yet", "verification": "", "reached": "", "outlook": "No credible window", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] } ] }, { "id": "energy", "name": "Energy", "short": "Energy", "stage": 1, "latency_log10_hours": 4.85, "latency": "Years", "latency_detail": "4 to 15 years: build, connect to the grid, switch on", "tracks": "AI in fusion control, grid planning, new energy materials.", "bottom_line": "AI helps run plasma experiments and grid queues. It hasn't changed how much energy gets built.", "held_back": "Power plants and grid hookups take 4 to 15 years.", "score": 30, "verified_score": 10, "previous_score": 30, "previous_verified_score": 10, "milestones": [ { "field_id": "energy", "n": 1, "milestone": "AI controls the plasma in a real fusion reactor", "status": "achieved", "verification": "peer_reviewed", "reached": "2022-02-16", "outlook": "", "evidence": "TCV tokamak, Nature", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [], "states": [ [ "2026-08-20", "achieved", "peer_reviewed" ] ] }, { "field_id": "energy", "n": 2, "milestone": "AI speeds up an energy bottleneck at utility scale", "status": "achieved", "verification": "claimed", "reached": "2026-04", "outlook": "", "evidence": "PJM: 811 grid applications sorted in under an hour", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "achieved", "claimed" ] ] }, { "field_id": "energy", "n": 3, "milestone": "An AI-found battery material works in a real cell", "status": "achieved", "verification": "preprint", "reached": "2024-01-09", "outlook": "", "evidence": "Microsoft and PNNL: 70% less lithium, lab prototype", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "achieved", "preprint" ] ] }, { "field_id": "energy", "n": 4, "milestone": "AI controls a net-energy fusion device", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2027 to 2029", "evidence": "SPARC not running yet", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "cfs" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "energy", "n": 5, "milestone": "An AI-found design improves a working energy asset", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2031", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "energy", "n": 6, "milestone": "An AI-found battery or solar chemistry in mass production", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2029 to 2033", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "energy", "n": 7, "milestone": "AI makes binding grid decisions", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2029 to 2032", "evidence": "Texas' 2026 audit uses human engineers", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "ercot_b0" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "energy", "n": 8, "milestone": "Fusion power on the grid, with AI in the controls", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2030 to 2034", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "energy", "n": 9, "milestone": "Energy stops limiting compute", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2032+", "evidence": "Texas alone has 474 GW of large-load requests", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "ercot_q" ], "states": [ [ "2026-08-20", "not_yet", "" ] ] }, { "field_id": "energy", "n": 10, "milestone": "AI-designed energy tech powers AI itself", "status": "not_yet", "verification": "", "reached": "", "outlook": "After 2035", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [], "states": [ [ "2026-08-20", "not_yet", "" ] ] } ] } ], "history": [ { "month": "2020-01", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-02", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-03", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-04", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-05", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-06", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-07", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-08", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-09", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-10", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-11", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2020-12", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2021-01", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2021-02", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2021-03", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2021-04", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2021-05", "index": 1.0, "verified": 0.0, "reached": 1 }, { "month": "2021-06", "index": 1.5, "verified": 0.5, "reached": 2 }, { "month": "2021-07", "index": 2.5, "verified": 1.5, "reached": 3 }, { "month": "2021-08", "index": 2.5, "verified": 1.5, "reached": 3 }, { "month": "2021-09", "index": 2.5, "verified": 1.5, "reached": 3 }, { "month": "2021-10", "index": 2.5, "verified": 1.5, "reached": 3 }, { "month": "2021-11", "index": 2.5, "verified": 1.5, "reached": 3 }, { "month": "2021-12", "index": 2.5, "verified": 1.5, "reached": 3 }, { "month": "2022-01", "index": 2.5, "verified": 1.5, "reached": 3 }, { "month": "2022-02", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2022-03", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2022-04", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2022-05", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2022-06", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2022-07", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2022-08", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2022-09", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2022-10", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2022-11", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2022-12", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2023-01", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2023-02", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2023-03", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2023-04", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2023-05", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2023-06", "index": 3.5, "verified": 2.5, "reached": 4 }, { "month": "2023-07", "index": 4.5, "verified": 3.5, "reached": 5 }, { "month": "2023-08", "index": 5.5, "verified": 4.5, "reached": 6 }, { "month": "2023-09", "index": 5.5, "verified": 4.5, "reached": 6 }, { "month": "2023-10", "index": 5.5, "verified": 4.5, "reached": 6 }, { "month": "2023-11", "index": 6.5, "verified": 5.5, "reached": 7 }, { "month": "2023-12", "index": 6.5, "verified": 5.5, "reached": 7 }, { "month": "2024-01", "index": 7.5, "verified": 5.5, "reached": 8 }, { "month": "2024-02", "index": 7.5, "verified": 5.5, "reached": 8 }, { "month": "2024-03", "index": 8.5, "verified": 5.5, "reached": 9 }, { "month": "2024-04", "index": 8.5, "verified": 5.5, "reached": 9 }, { "month": "2024-05", "index": 8.5, "verified": 5.5, "reached": 9 }, { "month": "2024-06", "index": 8.5, "verified": 5.5, "reached": 9 }, { "month": "2024-07", "index": 8.5, "verified": 5.5, "reached": 9 }, { "month": "2024-08", "index": 8.5, "verified": 5.5, "reached": 9 }, { "month": "2024-09", "index": 8.5, "verified": 5.5, "reached": 9 }, { "month": "2024-10", "index": 8.5, "verified": 5.5, "reached": 9 }, { "month": "2024-11", "index": 8.5, "verified": 5.5, "reached": 9 }, { "month": "2024-12", "index": 11.5, "verified": 7.5, "reached": 12 }, { "month": "2025-01", "index": 12.5, "verified": 8.5, "reached": 13 }, { "month": "2025-02", "index": 12.5, "verified": 8.5, "reached": 13 }, { "month": "2025-03", "index": 12.5, "verified": 8.5, "reached": 13 }, { "month": "2025-04", "index": 14.5, "verified": 9.5, "reached": 15 }, { "month": "2025-05", "index": 17.5, "verified": 9.5, "reached": 18 }, { "month": "2025-06", "index": 17.5, "verified": 9.5, "reached": 18 }, { "month": "2025-07", "index": 21.5, "verified": 13.5, "reached": 22 }, { "month": "2025-08", "index": 22.5, "verified": 14.5, "reached": 23 }, { "month": "2025-09", "index": 26.5, "verified": 16.5, "reached": 27 }, { "month": "2025-10", "index": 27.5, "verified": 17.5, "reached": 28 }, { "month": "2025-11", "index": 28.5, "verified": 18.5, "reached": 29 }, { "month": "2025-12", "index": 28.5, "verified": 18.5, "reached": 29 }, { "month": "2026-01", "index": 31.0, "verified": 19.5, "reached": 32 }, { "month": "2026-02", "index": 33.0, "verified": 19.5, "reached": 34 }, { "month": "2026-03", "index": 36.0, "verified": 20.5, "reached": 37 }, { "month": "2026-04", "index": 37.0, "verified": 20.5, "reached": 38 }, { "month": "2026-05", "index": 40.0, "verified": 22.5, "reached": 41 }, { "month": "2026-06", "index": 43.5, "verified": 25.0, "reached": 45 }, { "month": "2026-07", "index": 49.5, "verified": 28.0, "reached": 51 }, { "month": "2026-08", "index": 51.5, "verified": 29.0, "reached": 53 }, { "month": "2026-09", "index": 52.0, "verified": 29.0, "reached": 54 }, { "month": "2026-10", "index": 52.0, "verified": 29.0, "reached": 54 } ], "abilities": [ { "id": "mathematics", "name": "Math", "covers": "Proofs and problem solving", "compared_with": "Mathematicians, competition champions, the whole math community", "score": 100.0, "verified_score": 100.0, "steps": [ { "ability_id": "mathematics", "step": 1, "milestone": "AI writes machine-checked proofs of known results at research scale", "ref": "math:2", "status": "achieved", "verification": "replicated", "reached": "2025-09", "outlook": "", "evidence": "Strong Prime Number Theorem formalized in Lean in three weeks", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [] }, { "ability_id": "mathematics", "step": 2, "milestone": "AI solves a genuinely open problem, with a machine-checked proof", "ref": "math:3", "status": "achieved", "verification": "replicated", "reached": "2026-01-07", "outlook": "", "evidence": "Erdős problem #728, Lean proof", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "erdos728" ] }, { "ability_id": "mathematics", "step": 3, "milestone": "AI disproves a famous conjecture, confirmed by senior mathematicians", "ref": "math:4", "status": "achieved", "verification": "replicated", "reached": "2026-05-20", "outlook": "", "evidence": "Unit distance conjecture (1946), checked by nine mathematicians", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "unit", "unit_oai" ] } ] }, { "id": "coding", "name": "Coding", "covers": "Writing and fixing real software", "compared_with": "Professional engineers, world-finals programmers, whole engineering teams", "score": 75.0, "verified_score": 33.3, "steps": [ { "ability_id": "coding", "step": 1, "milestone": "AI agents solve over half of real coding tasks (SWE-bench Verified)", "ref": "ai:1", "status": "achieved", "verification": "claimed", "reached": "2024-12", "outlook": "", "evidence": "o3 reported 71.7% (Dec 2024). Latest: 81.4%, Claude Opus 4.6 (Feb 2026). Lab-reported", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [] }, { "ability_id": "coding", "step": 2, "milestone": "AI beats every human at a world programming final", "ref": "", "status": "achieved", "verification": "replicated", "reached": "2026-07", "outlook": "", "evidence": "AI won both divisions of AtCoder's 2026 world finals, per AtCoder. Earlier: 12 of 12 at ICPC 2025", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 10, "verified_points": 10, "sources": [ "atcoder_wtf26", "openai_icpc" ] }, { "ability_id": "coding", "step": 3, "milestone": "AI builds production software that took a whole team", "ref": "", "status": "partial", "verification": "claimed", "reached": "2026-09", "outlook": "An outside-checked build expected 2027 to 2028", "evidence": "GitHub says agents wrote most of its 800,000-line Copilot runtime, with people steering", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 5, "verified_points": 0, "sources": [ "github_rust" ] } ] }, { "id": "science", "name": "Science", "covers": "Discovery in physics, biology, chemistry and materials", "compared_with": "Researchers, the best human designs, whole research fields", "score": 83.3, "verified_score": 83.3, "steps": [ { "ability_id": "science", "step": 1, "milestone": "AI matches expert physicists on a research-grade test", "ref": "physics:6", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2027 to 2029", "evidence": "CritPt 32.3%, up from 5.7% in 11 months", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "critpt" ] }, { "ability_id": "science", "step": 2, "milestone": "An AI-designed instrument beats the human design", "ref": "physics:3", "status": "achieved", "verification": "peer_reviewed", "reached": "2025-04-14", "outlook": "", "evidence": "Urania detector designs", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "urania" ] }, { "ability_id": "science", "step": 3, "milestone": "AI solves a problem a whole field couldn't for decades", "ref": "", "status": "achieved", "verification": "replicated", "reached": "2020-11-30", "outlook": "", "evidence": "AlphaFold 2: independent CASP assessors judged the 50-year protein folding problem largely solved, for single proteins", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 10, "verified_points": 10, "sources": [ "casp14", "alphafold2" ] } ] }, { "id": "medicine", "name": "Medicine", "covers": "Diagnosis and treatment", "compared_with": "Doctors, top specialists, medicine as a field", "score": 16.7, "verified_score": 16.7, "steps": [ { "ability_id": "medicine", "step": 1, "milestone": "AI matches doctors at diagnosis in a blinded study", "ref": "", "status": "achieved", "verification": "peer_reviewed", "reached": "2025-04-09", "outlook": "", "evidence": "Google's AMIE matched or beat 20 primary care doctors in blinded text consultations, with patient actors", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 10, "verified_points": 10, "sources": [ "amie" ] }, { "ability_id": "medicine", "step": 2, "milestone": "AI beats doctors on real patient outcomes, in a prospective trial", "ref": "bio:7", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2029 to 2033", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [] }, { "ability_id": "medicine", "step": 3, "milestone": "A disease durably cured by an AI-designed therapy", "ref": "bio:9", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2032 to 2035+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [] } ] }, { "id": "engineering", "name": "Engineering", "covers": "Designing chips, circuits and devices", "compared_with": "Design engineers, the best human designs, whole design teams", "score": 50.0, "verified_score": 33.3, "steps": [ { "ability_id": "engineering", "step": 1, "milestone": "An AI-written circuit change ships in a production accelerator", "ref": "chips:4", "status": "achieved", "verification": "claimed", "reached": "2025-05-14", "outlook": "", "evidence": "AlphaEvolve", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "alphaevolve" ] }, { "ability_id": "engineering", "step": 2, "milestone": "An AI-designed radio-frequency block beats human designs when built", "ref": "chips:5", "status": "achieved", "verification": "peer_reviewed", "reached": "2024-12-30", "outlook": "", "evidence": "Designed in minutes rather than weeks", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 10, "sources": [ "rfblock" ] }, { "ability_id": "engineering", "step": 3, "milestone": "A full production chip designed with no human intervention", "ref": "chips:7", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2030 to 2034", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [ "semieng" ] } ] }, { "id": "work", "name": "Professional work", "covers": "The everyday work of skilled jobs", "compared_with": "Experts in 44 occupations, top experts, whole companies", "score": 50.0, "verified_score": 0.0, "steps": [ { "ability_id": "work", "step": 1, "milestone": "AI matches experts on half of real work tasks", "ref": "econ:3", "status": "achieved", "verification": "preprint", "reached": "2025-09-25", "outlook": "", "evidence": "GDPval, 44 occupations, blind grading", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [] }, { "ability_id": "work", "step": 2, "milestone": "AI's work is preferred to experts' on most real tasks", "ref": "", "status": "achieved", "verification": "claimed", "reached": "2025-12-11", "outlook": "", "evidence": "GPT-5.2 Pro's work beat experts' outright in 60% of blind GDPval comparisons", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 10, "verified_points": 0, "sources": [ "gpt52" ] }, { "ability_id": "work", "step": 3, "milestone": "An AI-run business outperforms human-run rivals", "ref": "", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2028 to 2031", "evidence": "An AI-run San Francisco shop, open since April 2026, is not yet profitable", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 0, "verified_points": 0, "sources": [ "andon_market" ] } ] }, { "id": "physical", "name": "Physical skill", "covers": "Driving, flying and handling the real world", "compared_with": "Professional drivers, world champions, whole workforces", "score": 50.0, "verified_score": 33.3, "steps": [ { "ability_id": "physical", "step": 1, "milestone": "A driverless fleet proves safer than humans over 100M miles", "ref": "robo:4", "status": "achieved", "verification": "claimed", "reached": "2026-03", "outlook": "", "evidence": "Waymo: 220.6M miles, 94% fewer serious-injury crashes. Company data", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 10, "verified_points": 0, "sources": [ "waymo" ] }, { "ability_id": "physical", "step": 2, "milestone": "A robot beats human champions at a physical sport", "ref": "", "status": "achieved", "verification": "peer_reviewed", "reached": "2023-08-30", "outlook": "", "evidence": "Swift, an autonomous racing drone, won 15 of 25 races against three champion pilots, on one track", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 10, "verified_points": 10, "sources": [ "swift" ] }, { "ability_id": "physical", "step": 3, "milestone": "Physical labor automated at economic scale", "ref": "robo:9", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2032 to 2035+", "evidence": "", "recorded": "2026-08-20", "updated": "2026-08-20", "points": 0, "verified_points": 0, "sources": [] } ] }, { "id": "forecasting", "name": "Forecasting", "covers": "Judging what will happen", "compared_with": "The public, superforecasters, crowds and markets", "score": 33.3, "verified_score": 33.3, "steps": [ { "ability_id": "forecasting", "step": 1, "milestone": "AI forecasts as well as the public", "ref": "", "status": "achieved", "verification": "replicated", "reached": "2025-10-08", "outlook": "", "evidence": "ForecastBench: the best AI system scores 69.7 against the public's 63.3", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 10, "verified_points": 10, "sources": [ "forecastbench", "fri_public" ] }, { "ability_id": "forecasting", "step": 2, "milestone": "AI beats superforecasters", "ref": "", "status": "partial", "verification": "replicated", "reached": "2026-07-16", "outlook": "A clear win expected 2027 to 2028", "evidence": "ForecastBench: best AI 69.7 against superforecasters' 68.5, a statistical tie", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 5, "verified_points": 5, "sources": [ "forecastbench", "fri_parity" ] }, { "ability_id": "forecasting", "step": 3, "milestone": "AI beats superforecasters and prediction markets, without seeing their forecasts", "ref": "", "status": "not_yet", "verification": "", "reached": "", "outlook": "Expected 2027 to 2029", "evidence": "AI that can't see market prices still trails them: 0.126 against 0.111 (lower is better)", "recorded": "2026-10-29", "updated": "2026-10-29", "points": 0, "verified_points": 0, "sources": [ "aia_forecaster" ] } ] } ], "stories": [ { "date": "2026-10-08", "field_id": "math", "verification": "peer_reviewed", "headline": "A top journal publishes AI-made math", "summary": "Science published DeepMind's AlphaProof Nexus on 8 Oct. It settled 9 Erdős problems.", "sources": [ "nexus", "unit_pnas" ] }, { "date": "2026-09-08", "field_id": "math", "verification": "claimed", "headline": "OpenAI says it solved Navier–Stokes", "summary": "A Lean proof of one version of the Millennium Prize problem, posted 8 Sep. No referee or Clay ruling yet, so it counts as half a milestone.", "sources": [ "navier" ] }, { "date": "2026-10-06", "field_id": "math", "verification": "claimed", "headline": "721 AI math papers in one release", "summary": "OpenAI posted them on 6 Oct, about 42% with machine-checked proofs. It withdrew three the next day.", "sources": [ "oai_bulk", "oai_hist" ] }, { "date": "2026-10", "field_id": "math", "verification": "replicated", "headline": "A Claude proof, rebuilt by hand", "summary": "Claude's Lean proof on critical percolation has been re-derived by two outside mathematicians.", "sources": [ "perc", "perc_br", "perc_g" ] }, { "date": "2026-10", "field_id": "math", "verification": "data", "headline": "10 of 49 open problems solved", "summary": "Epoch AI's open-problems tracker, up from 4 in August. Half had human help.", "sources": [ "fm_open" ] }, { "date": "2026-10-29", "field_id": "ai", "verification": "data", "headline": "Reliability gap: 5.6 to 1", "summary": "Top models finish 17.4-hour tasks half the time, but only 3.1-hour tasks reliably (80% of the time).", "sources": [ "metr" ] }, { "date": "2026-10-29", "field_id": "bio", "verification": "data", "headline": "Still no AI-designed drug approved", "summary": "The first FDA decision to watch is zasocitinib, due Q1 2027. Rentosertib's Phase III began dosing on 10 Sep.", "sources": [ "fda_novel", "zaso", "rento_dose" ] }, { "date": "2026-09-11", "field_id": "energy", "verification": "data", "headline": "Texas pauses new data center hookups", "summary": "An audit of the grid's large-load requests, mostly data centers. Reports due 10 Dec.", "sources": [ "tx_gov" ] }, { "date": "2026-10", "field_id": "econ", "verification": "data", "headline": "24.6% of US firms now use AI", "summary": "Census survey, two weeks to 4 Oct. Up from 21.8% in July.", "sources": [ "btos" ] } ], "metrics": [ { "metric_id": "metr_80_ratio", "date": "2026-04", "value": 5.6, "source_id": "metr", "note": "17.4 h at 50% success vs 3.1 h at 80% (early Claude Mythos Preview)" }, { "metric_id": "btos_ai_use", "date": "2026-07-26", "value": 21.8, "source_id": "btos", "note": "Two weeks ending 26 Jul 2026" }, { "metric_id": "btos_ai_use", "date": "2026-10-04", "value": 24.6, "source_id": "btos", "note": "Two weeks ending 4 Oct 2026" }, { "metric_id": "fm_open_solved", "date": "2026-08", "value": 4.0, "source_id": "fm_open", "note": "Of 49 listed open problems" }, { "metric_id": "fm_open_solved", "date": "2026-10", "value": 10.0, "source_id": "fm_open", "note": "Of 49 listed; 5 by AI alone, 5 by humans with AI" }, { "metric_id": "us_tfp_growth", "date": "2025", "value": 0.8, "source_id": "bls", "note": "US total factor productivity growth, 2025" }, { "metric_id": "tx_large_load_gw", "date": "2026-06", "value": 474.0, "source_id": "ercot_q", "note": "ERCOT large-load requests of 75 MW or more" }, { "metric_id": "ai_drug_approvals", "date": "2026-10-29", "value": 0.0, "source_id": "fda_novel", "note": "FDA or EMA approvals of drugs credited to AI design" } ], "watch": [ { "metric_id": "metr_80_ratio", "unit": "ratio", "title": "Reliability gap", "why": "The length of task AI finishes half the time, divided by the length it finishes 8 times in 10. Past 10 to 1, AI is getting faster than it is getting reliable." }, { "metric_id": "btos_ai_use", "unit": "%", "title": "US firms using AI", "why": "Share of US businesses that used AI in the last two weeks, from the Census Bureau. The survey changed in Nov 2025, so older readings don't compare." }, { "metric_id": "fm_open_solved", "unit": "of 49", "title": "Open math problems solved", "why": "Problems on Epoch AI's list of open math problems that AI has solved, alone or with people. Fewer than 15 by Aug 2027 would mean progress has stalled." }, { "metric_id": "us_tfp_growth", "unit": "%", "title": "US productivity growth", "why": "How much more the US economy gets from the same work and capital each year. Two years above 1.5% would be the first sign of AI lifting the economy. Next reading Mar 2027." }, { "metric_id": "tx_large_load_gw", "unit": "GW", "title": "Texas large-load power requests", "why": "Power that data centers and other large users have asked the Texas grid to supply. Compare it with power actually switched on: a wide gap points to overbuilding." } ], "gatekeepers": [ { "id": "math_journals", "name": "Math journals", "signed_off": "yes", "since": "2026-05", "note": "Accepting AI-proved papers since May 2026, starting with Axiom Math.", "sources": [ "nexus", "axiom1", "axiom5" ] }, { "id": "drug_regulators", "name": "Drug regulators", "signed_off": "no", "since": "", "note": "No AI-designed drug approved by the FDA or EMA. First decision to watch: zasocitinib, due Q1 2027, though it was designed with physics-based software rather than generative AI. Rentosertib's Phase III began dosing on 10 Sep.", "sources": [ "fda_novel", "zaso", "rento_dose" ] }, { "id": "standards_bodies", "name": "Standards bodies", "signed_off": "no", "since": "", "note": "No standards body has adopted a result that AI originated.", "sources": [] } ], "updates": [ { "date": "2026-08-20", "edition": "v1.0", "summary": "First edition. Ten fields, 100 milestones." }, { "date": "2026-10-29", "edition": "v1.1", "summary": "Math rises from 75 to 85. Every result is now labeled by who checked it." } ], "corrections": [ { "date": "2026-10-29", "field_id": "chips", "n": "", "was": "Chips scored 50%", "now": "45%: the contested AlphaChip milestone was shown partial but scored achieved" }, { "date": "2026-10-29", "field_id": "", "n": "", "was": "Zero AI results ratified by a journal, regulator or standards body", "now": "Restated as a gatekeeper count: math journals yes, regulators and standards bodies no" }, { "date": "2026-10-29", "field_id": "ai", "n": "", "was": "METR publishes no 80% horizons", "now": "They are in METR's data file: 17.4 h at 50%, 3.1 h at 80%" }, { "date": "2026-10-29", "field_id": "math", "n": "8", "was": "Axiom's journal acceptances unconfirmed", "now": "Confirmed: five journals, five DOIs" }, { "date": "2026-10-29", "field_id": "energy", "n": "", "was": "Texas stopped accepting new data center development", "now": "It ordered an audit and paused energization approvals" }, { "date": "2026-10-29", "field_id": "econ", "n": "1", "was": "Census BTOS 21.8%, May to July", "now": "A single two-week reading ending 26 Jul 2026" }, { "date": "2026-10-29", "field_id": "bio", "n": "", "was": "173 AI drug programs in trials", "now": "Untraceable; replaced with 117 (to Jul 2025) and 129 (to Mar 2026)" }, { "date": "2026-10-29", "field_id": "bio", "n": "", "was": "AI drugs succeed in Phase II at about 40%", "now": "A 2026 analysis puts it near 20%" }, { "date": "2026-10-29", "field_id": "chips", "n": "1", "was": "Neither Synopsys nor Cadence publishes tapeout counts", "now": "Both do; unaudited" }, { "date": "2026-10-29", "field_id": "mat", "n": "10", "was": "Alibaba published no compounds or temperatures", "now": "Four rhenium compounds, 2.5 to 6.5 K, in a preprint" }, { "date": "2026-10-29", "field_id": "bio", "n": "", "was": "Missed Science's AI-generated phage genomes (6 Aug 2026)", "now": "Added" } ], "sources": [ { "id": "erdos728", "label": "arXiv:2601.07421", "url": "https://arxiv.org/abs/2601.07421", "kind": "preprint" }, { "id": "unit", "label": "arXiv:2605.20695", "url": "https://arxiv.org/abs/2605.20695", "kind": "preprint" }, { "id": "unit_oai", "label": "OpenAI", "url": "https://openai.com/index/model-disproves-discrete-geometry-conjecture/", "kind": "company" }, { "id": "unit_pnas", "label": "PNAS", "url": "https://doi.org/10.1073/pnas.2621541123", "kind": "journal" }, { "id": "jacobian", "label": "Tao", "url": "https://terrytao.wordpress.com/2026/07/21/a-digestion-of-the-jacobian-conjecture-counterexample/", "kind": "commentary" }, { "id": "jacobian_afp", "label": "Isabelle AFP", "url": "https://www.isa-afp.org/entries/Jacobian_Counterexample.html", "kind": "data" }, { "id": "astra_gh", "label": "Lean repo", "url": "https://github.com/openai/ten-proofs", "kind": "code" }, { "id": "riemann", "label": "Anthropic", "url": "https://www.anthropic.com/research/riemann-zeta", "kind": "company" }, { "id": "riemann_paper", "label": "arXiv:2608.13637", "url": "https://arxiv.org/abs/2608.13637", "kind": "preprint" }, { "id": "lamzouri", "label": "arXiv:2609.02882", "url": "https://arxiv.org/abs/2609.02882", "kind": "preprint" }, { "id": "nexus", "label": "Science", "url": "https://doi.org/10.1126/science.aej2213", "kind": "journal" }, { "id": "axiom1", "label": "Indag. Math.", "url": "https://doi.org/10.1016/j.indag.2026.05.002", "kind": "journal" }, { "id": "axiom5", "label": "J. Combin. Theory A", "url": "https://doi.org/10.1016/j.jcta.2026.106237", "kind": "journal" }, { "id": "navier", "label": "OpenAI", "url": "https://openai.com/index/navier-stokes-solution/", "kind": "company" }, { "id": "oai_bulk", "label": "OpenAI", "url": "https://openai.com/index/sharing-ai-progress-in-mathematics/", "kind": "company" }, { "id": "oai_repo", "label": "openai/math", "url": "https://github.com/openai/math", "kind": "code" }, { "id": "oai_hist", "label": "Release history", "url": "https://github.com/openai/math/blob/main/history.md", "kind": "code" }, { "id": "oai_oct26", "label": "OpenAI", "url": "https://openai.com/index/building-scientific-understanding-with-academic-communities", "kind": "company" }, { "id": "perc", "label": "Lean proof", "url": "https://github.com/anthropics/formal-math/tree/795efb86f191735c5481675763537cfb4ff37e55/percolation", "kind": "code" }, { "id": "perc_br", "label": "arXiv:2610.10502", "url": "https://arxiv.org/abs/2610.10502", "kind": "preprint" }, { "id": "perc_g", "label": "arXiv:2610.30063", "url": "https://arxiv.org/abs/2610.30063", "kind": "preprint" }, { "id": "fm_open", "label": "Epoch AI", "url": "https://epoch.ai/frontiermath/open-problems", "kind": "data" }, { "id": "metr", "label": "METR data", "url": "https://metr.org/assets/benchmark_results_1_1.yaml", "kind": "data" }, { "id": "discorl", "label": "Nature", "url": "https://www.nature.com/articles/s41586-025-09761-x", "kind": "journal" }, { "id": "alphaevolve", "label": "DeepMind", "url": "https://deepmind.google/discover/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/", "kind": "company" }, { "id": "sol", "label": "OpenAI", "url": "https://openai.com/index/gpt-5-6-frontier-intelligence-efficiency/", "kind": "company" }, { "id": "rsi", "label": "Anthropic", "url": "https://www.anthropic.com/institute/recursive-self-improvement", "kind": "company" }, { "id": "w2s", "label": "Anthropic", "url": "https://alignment.anthropic.com/2026/automated-w2s-researcher/", "kind": "company" }, { "id": "ligo", "label": "Science", "url": "https://www.science.org/doi/10.1126/science.adw1291", "kind": "journal" }, { "id": "urania", "label": "Phys. Rev. X", "url": "https://journals.aps.org/prx/abstract/10.1103/PhysRevX.15.021012", "kind": "journal" }, { "id": "plasma", "label": "PNAS", "url": "https://www.pnas.org/doi/10.1073/pnas.2505725122", "kind": "journal" }, { "id": "euler", "label": "DeepMind", "url": "https://deepmind.google/blog/discovering-new-solutions-to-century-old-problems-in-fluid-dynamics/", "kind": "company" }, { "id": "critpt", "label": "Artificial Analysis", "url": "https://artificialanalysis.ai/evaluations/critpt", "kind": "data" }, { "id": "scet", "label": "arXiv:2601.02484", "url": "https://arxiv.org/abs/2601.02484", "kind": "preprint" }, { "id": "grad", "label": "arXiv:2609.26742", "url": "https://arxiv.org/abs/2609.26742", "kind": "preprint" }, { "id": "grad_umd", "label": "UMD", "url": "https://ece.umd.edu/news/story/59year-old-mathematical-conjecture-disproved-by-ece-professor", "kind": "company" }, { "id": "epoch_price", "label": "Epoch AI", "url": "https://epoch.ai/data-insights/llm-inference-price-trends", "kind": "data" }, { "id": "alphachip", "label": "DeepMind", "url": "https://deepmind.google/discover/blog/how-alphachip-transformed-computer-chip-design/", "kind": "company" }, { "id": "rfblock", "label": "Nature Communications", "url": "https://www.nature.com/articles/s41467-024-54178-1", "kind": "journal" }, { "id": "semieng", "label": "Semiconductor Engineering", "url": "https://semiengineering.com/the-impact-of-ai-automation-on-chip-design/", "kind": "commentary" }, { "id": "synopsys", "label": "Synopsys", "url": "https://investor.synopsys.com/news/news-details/2024/Synopsys-Announces-New-AI-Driven-EDA-IP-and-Systems-Design-Solutions-At-SNUG-Silicon-Valley/default.aspx", "kind": "company" }, { "id": "cadence", "label": "Cadence", "url": "https://www.cadence.com/en_US/home/company/newsroom/press-releases/pr/2026/cadence-unleashes-chipstack-ai-super-agent-pioneering-a-new.html", "kind": "company" }, { "id": "protein", "label": "Nature", "url": "https://www.nature.com/articles/s41586-025-09721-5", "kind": "journal" }, { "id": "rento", "label": "Insilico", "url": "https://insilico.com/news/xmjsn4l091-insilico-initiates-phase-iii-clinical-tr", "kind": "company" }, { "id": "rento_p2", "label": "Nature Medicine", "url": "https://doi.org/10.1038/s41591-025-03743-2", "kind": "journal" }, { "id": "rento_dose", "label": "Insilico", "url": "https://insilico.com/news/isn1009261-insilico-medicine-doses-first-patient-genesis-ipf-3", "kind": "company" }, { "id": "zaso", "label": "Takeda", "url": "https://www.takeda.com/newsroom/newsreleases/2026/fda-priority-review-zasocitinib-psoriasis/", "kind": "company" }, { "id": "fda_novel", "label": "FDA", "url": "https://www.fda.gov/drugs/novel-drug-approvals-fda/novel-drug-approvals-2026", "kind": "government" }, { "id": "evo", "label": "Science", "url": "https://www.science.org/doi/10.1126/science.aec2657", "kind": "journal" }, { "id": "mattergen", "label": "Nature", "url": "https://www.nature.com/articles/s41586-025-08628-5", "kind": "journal" }, { "id": "mnfeco", "label": "arXiv:2603.22748", "url": "https://arxiv.org/abs/2603.22748", "kind": "preprint" }, { "id": "kagome", "label": "Rice", "url": "https://www.eurekalert.org/news-releases/1133828", "kind": "commentary" }, { "id": "catalyst", "label": "Nature Communications", "url": "https://www.nature.com/articles/s41467-026-74425-x", "kind": "journal" }, { "id": "gnome_dup", "label": "arXiv:2610.13885", "url": "https://arxiv.org/abs/2610.13885", "kind": "preprint" }, { "id": "lila", "label": "arXiv:2609.30133", "url": "https://doi.org/10.48550/arxiv.2609.30133", "kind": "preprint" }, { "id": "dmat", "label": "TechCrunch", "url": "https://techcrunch.com/2026/08/10/discovered-materials-is-playing-ai-whack-a-mole-to-hunt-cooler-chips/", "kind": "commentary" }, { "id": "elements", "label": "arXiv:2604.23758", "url": "https://arxiv.org/abs/2604.23758", "kind": "preprint" }, { "id": "gr2", "label": "DeepMind", "url": "https://deepmind.google/blog/gemini-robotics-2-brings-whole-body-intelligence-to-robots", "kind": "company" }, { "id": "waymo", "label": "Waymo", "url": "https://waymo.com/safety/impact/", "kind": "company" }, { "id": "tesla", "label": "Tesla", "url": "https://assets-ir.tesla.com/tesla-contents/IR/TSLA-Q3-2026-Update.pdf", "kind": "company" }, { "id": "btos", "label": "Census", "url": "https://www.census.gov/hfp/btos/downloads/National.xlsx", "kind": "government" }, { "id": "canaries", "label": "Stanford", "url": "https://digitaleconomy.stanford.edu/publications/canaries-in-the-coal-mine/", "kind": "company" }, { "id": "bls", "label": "BLS", "url": "https://www.bls.gov/news.release/prod3.nr0.htm", "kind": "government" }, { "id": "x402", "label": "Coinbase", "url": "https://www.coinbase.com/developer-platform/discover/launches/c4a-equities-x402", "kind": "company" }, { "id": "chainalysis", "label": "Chainalysis", "url": "https://www.chainalysis.com/blog/north-america-crypto-adoption-2026/", "kind": "company" }, { "id": "bci", "label": "Nature Medicine", "url": "https://www.nature.com/articles/s41591-026-04414-6", "kind": "journal" }, { "id": "paradromics", "label": "Paradromics", "url": "https://paradromics.com/news/paradromics-receives-fda-approval-to-expand-connexus-brain-computer-interface-clinical-study-recruitment-nationwide/", "kind": "company" }, { "id": "ercot_q", "label": "ERCOT", "url": "https://www.ercot.com/files/docs/2026/07/29/ERCOT-Senate-July-29-Panel-1-Assessing-The-Grid.pdf", "kind": "government" }, { "id": "ercot_b0", "label": "ERCOT", "url": "https://www.ercot.com/files/docs/2026/09/11/14-Batch-Zero-Update.pdf", "kind": "government" }, { "id": "tx_gov", "label": "Texas Governor", "url": "https://gov.texas.gov/news/post/governor-abbott-directs-comprehensive-data-center-audit", "kind": "government" }, { "id": "cfs", "label": "CFS", "url": "https://blog.cfs.energy/sparc-progress-doe-milestone-approvals-reflect-cfs-power-plant-maturity/", "kind": "company" }, { "id": "casp14", "label": "CASP14 assessment (Proteins)", "url": "https://doi.org/10.1002/prot.26237", "kind": "journal" }, { "id": "alphafold2", "label": "Nature (AlphaFold 2)", "url": "https://doi.org/10.1038/s41586-021-03819-2", "kind": "journal" }, { "id": "amie", "label": "Nature (AMIE)", "url": "https://doi.org/10.1038/s41586-025-08866-7", "kind": "journal" }, { "id": "swift", "label": "Nature (Swift)", "url": "https://doi.org/10.1038/s41586-023-06419-4", "kind": "journal" }, { "id": "atcoder_wtf26", "label": "AtCoder", "url": "https://info.atcoder.jp/entry/awtf2026-report", "kind": "data" }, { "id": "openai_icpc", "label": "OpenAI on GitHub", "url": "https://github.com/openai/openai-icpc-2025", "kind": "code" }, { "id": "github_rust", "label": "GitHub", "url": "https://github.blog/ai-and-ml/generative-ai/migrating-the-github-copilot-runtime-to-rust-using-copilot/", "kind": "company" }, { "id": "gpt52", "label": "OpenAI", "url": "https://openai.com/index/introducing-gpt-5-2/", "kind": "company" }, { "id": "andon_market", "label": "Andon Labs", "url": "https://andonlabs.com/market", "kind": "company" }, { "id": "forecastbench", "label": "ForecastBench", "url": "https://forecastbench.org/leaderboards/", "kind": "data" }, { "id": "fri_public", "label": "Forecasting Research Institute", "url": "https://forecastingresearch.substack.com/p/ai-llm-forecasting-model-forecastbench-benchmark", "kind": "commentary" }, { "id": "fri_parity", "label": "Forecasting Research Institute", "url": "https://forecastingresearch.substack.com/p/ai-models-have-likely-reached-parity", "kind": "commentary" }, { "id": "aia_forecaster", "label": "arXiv:2511.07678", "url": "https://arxiv.org/abs/2511.07678", "kind": "preprint" } ] }