AI will allow humanity to complete the next 25 years of scientific discovery in just 5 years — effectively bringing 2050-era science to 2030. Already manifesting in resolution of long-standing mathematical proofs (e.g. Erdős problems) via advanced reas...
Predictor: Kevin Weil
Prediction text
AI will allow humanity to complete the next 25 years of scientific discovery in just 5 years — effectively bringing 2050-era science to 2030. Already manifesting in resolution of long-standing mathematical proofs (e.g. Erdős problems) via advanced reasoning models. | First Nobel Prize awarded to AI-coauthored research
Key catalyst: First Nobel Prize awarded to AI-coauthored research
Watch events: Nature / Science AI-coauthored breakthroughs; Clay-Institute problems progress
Resolution evidence
Erdős-problem proofs resolved via reasoning models 2025; AlphaProof / AlphaGeometry / OpenAI o3 mathematical reasoning advances 2024-2026.
Predictor: Kevin Weil
Calibration plot (stated vs observed)
Evidence about this node from Kevin Weil is multiplied by κ in /api/intake. Lower κ = less weight; floors at 0.10 (effectively silenced) and caps at 1.00 (full weight).
Reference class
This node isn't linked to a reference class. The Bayesian update applies without outside-view blending.
Probability over time
Milestone chain
- 2026-01-08hitGPT-5.2 Pro autonomously solves Erdos Problem #728 (Tao verified)How: Terence Tao publicly confirms GPT-5.2 generated original proof of Erdos #728, formalized in Lean via AristotleSource: https://www.scientificamerican.com/article/ai-uncovers-solutions-to-erdos-problems-moving-closer-to-transforming-math/conf 99%Notes: HIT — first AI-coauthored peer-validatable original mathematical result. Direct match to ROB_003 narrative.
- 2026-01-15hitAI cracks 3+ Erdos problems within 7-day window (Jan 2026)How: GPT-5.2 + Aristotle/Lean stack solves Erdos #728, #729, and #397 within ~7-day periodSource: https://medium.com/@cognidownunder/three-erds-problems-fell-in-seven-days-and-terence-tao-verified-every-proof-himself-1a1ff4399bc6conf 97%
- 2026-11-24pendingQ1 window check-in (25%)
- 2026-06-01 → 2027-12-31pendingAI-coauthored result accepted in major peer-reviewed math journalHow: Annals of Mathematics / Inventiones / JAMS / similar accepts paper with AI listed as proof contributorSource: MathSciNet, journal acceptance announcementsconf 65%
- 2027-10-17pendingQ2 window check-in (50%)
- 2028-09-08pendingQ3 window check-in (75%)
- 2027-10-01 → 2030-10-31pendingFirst Nobel Prize awarded to AI-coauthored researchHow: Nobel Prize (Chemistry, Physics, Medicine, Economics) explicitly cites AI system as material contributor to award-winning researchSource: Nobel Prize committee announcements (Oct annually)conf 40%Notes: Listed as anchor in original prediction description. AlphaFold 2024 Chemistry Nobel was a partial precedent.
- 2028-06-01 → 2030-08-31pendingMajor scientific organization (NSF/Royal Society/AAAS) declares 5x research acceleration achievedHow: NSF, Royal Society, AAAS, or equivalent publishes report concluding AI-accelerated research compresses 10+ years of progress into 5 or fewer years in at least one major scientific domainSource: NSF science indicators, Royal Society annual reportsconf 45%
No downstream cascades — this prediction is a leaf in the dependency graph.
What if this resolves?
Click a button to clamp this prediction and run a Gibbs sample. Returns the predictions whose marginals shift most. ~30s per run; ideal for stress-testing "if X resolves, what else moves?"
Evidence chain
Network propagation neighbors
No propagation data yet. Run inference/.venv/bin/python scripts/ops/run_loopy_belief_propagation.py on the droplet, or wait for the Sunday 02:00 UTC weekly cron.
Prerequisites (2)
| Type | Pred | Title | Domain | Lag |
|---|---|---|---|---|
| correlate | S_AGI_MID_2029 | AGI mid: Kurzweil 2029 path | agi_general_capability | — |
| correlate | S_AGI_SLOW_2031 | AGI slow: Schmidt/Hassabis 5-10 year path | agi_general_capability | — |
Dependents (0)
| Type | Pred | Title | Domain | Lag |
|---|---|---|---|---|
| No dependents | ||||
Validations (1)
| Observed at | Status | By | Notes |
|---|---|---|---|
| 2026-04-29 | partial | thesis_timeline_v1.0_import | Erdős-problem proofs resolved via reasoning models 2025; AlphaProof / AlphaGeometry / OpenAI o3 mathematical reasoning advances 2024-2026. |
Linked documents (10)
Raw metadata
{
"nia": false,
"qty": "25 years in 5 years (5x)",
"mode": "FORECAST",
"role": "Cited-Other",
"context": "Specific 5x compression framing distinct from AI_031 (Hassabis '100 years of biology in a fraction of time'). Couples with SPC_022 ('2026 is to AI-science what 2025 was to AI-software').",
"to_year": 2030,
"conv_cues": "explicit compression ratio; named evidentiary milestones",
"direction": "HAPPEN",
"from_year": 2026,
"timeframe": "2026-2030",
"conv_level": "HIGH",
"milestones": [
{
"kind": "llm_pre_event",
"label": "GPT-5.2 Pro autonomously solves Erdos Problem #728 (Tao verified)",
"notes": "HIT — first AI-coauthored peer-validatable original mathematical result. Direct match to ROB_003 narrative.",
"source": "https://www.scientificamerican.com/article/ai-uncovers-solutions-to-erdos-problems-moving-closer-to-transforming-math/",
"status": "hit",
"weight": 0.4,
"ordinal": -9,
"source_id": null,
"confidence": 0.99,
"source_url": "https://www.scientificamerican.com/article/ai-uncovers-solutions-to-erdos-problems-moving-closer-to-transforming-math/",
"expected_date": "2026-01-08",
"observed_date": "2026-01-08",
"research_origin": "deep_research",
"measurement_criterion": "Terence Tao publicly confirms GPT-5.2 generated original proof of Erdos #728, formalized in Lean via Aristotle"
},
{
"kind": "llm_pre_event",
"label": "AI cracks 3+ Erdos problems within 7-day window (Jan 2026)",
"source": "https://medium.com/@cognidownunder/three-erds-problems-fell-in-seven-days-and-terence-tao-verified-every-proof-himself-1a1ff4399bc6",
"status": "hit",
"weight": 0.4,
"ordinal": -8,
"source_id": null,
"confidence": 0.97,
"source_url": "https://medium.com/@cognidownunder/three-erds-problems-fell-in-seven-days-and-terence-tao-verified-every-proof-himself-1a1ff4399bc6",
"expected_date": "2026-01-15",
"observed_date": "2026-01-15",
"research_origin": "deep_research",
"measurement_criterion": "GPT-5.2 + Aristotle/Lean stack solves Erdos #728, #729, and #397 within ~7-day period"
},
{
"kind": "quartile_checkpoint",
"label": "Q1 window check-in (25%)",
"status": "pending",
"weight": 0.05,
"ordinal": -7,
"source_id": null,
"expected_date": "2026-11-24",
"observed_date": null
},
{
"kind": "llm_pre_event",
"label": "AI-coauthored result accepted in major peer-reviewed math journal",
"source": "MathSciNet, journal acceptance announcements",
"status": "pending",
"weight": 0.4,
"ordinal": -6,
"source_id": null,
"confidence": 0.65,
"expected_date": "2027-03-17",
"research_origin": "deep_research",
"expected_date_range": {
"to": "2027-12-31",
"from": "2026-06-01"
},
"measurement_criterion": "Annals of Mathematics / Inventiones / JAMS / similar accepts paper with AI listed as proof contributor"
},
{
"kind": "quartile_checkpoint",
"label": "Q2 window check-in (50%)",
"status": "pending",
"weight": 0.05,
"ordinal": -5,
"source_id": null,
"expected_date": "2027-10-17",
"observed_date": null
},
{
"kind": "quartile_checkpoint",
"label": "Q3 window check-in (75%)",
"status": "pending",
"weight": 0.05,
"ordinal": -4,
"source_id": null,
"expected_date": "2028-09-08",
"observed_date": null
},
{
"kind": "scenario_signal",
"label": "Scenario fires: AGI mid: Kurzweil 2029 path",
"status": "pending",
"weight": 0.5,
"ordinal": -3,
"source_id": "S_AGI_MID_2029",
"expected_date": "2029-03-31",
"observed_date": null
},
{
"kind": "llm_post_event",
"label": "First Nobel Prize awarded to AI-coauthored research",
"notes": "Listed as a
... (truncated)