← Cockpit
INF_072predictionAIAGI-scaling-50-50

There is approximately a 50/50 chance that simply scaling existing methodologies (transformer architecture + more data + more compute) will be enough to reach AGI — though "nowhere near" human-level AGI currently.

Predictor: Demis Hassabis

Prior probability
50.0%
Current probability
40.4%
evolves via intake + LBP
Conviction
3/5
Signal quality
A
Resolution
pending
Window
2030-01-01 – 2042-09-30
Edges in / out
7 / 0
Tickers exposed
19

Prediction text

There is approximately a 50/50 chance that simply scaling existing methodologies (transformer architecture + more data + more compute) will be enough to reach AGI — though "nowhere near" human-level AGI currently. | Next DeepMind model-release capability evaluation

Key catalyst: Next DeepMind model-release capability evaluation

Watch events: GDPval-scale benchmarks; novel-architecture release from DeepMind

Resolution evidence

Status: pending

Hassabis position consistent across 2024-2026 interviews. AlphaFold Nobel (2024) reinforces his scientific-breakthrough-required thesis.

Predictor: Demis Hassabis

κ + Brier as of 2026-05-22
κ (discount)
0.583
Brier
0.0064
excellent
Hits / Misses
1 / 0
of 1 resolved
Hit rate
100.0%
Calibration plot (stated vs observed)

Evidence about this node from Demis Hassabis 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: agi_breakthrough_5y

Linked via embedding similarity 0.662

Major capability discontinuity (e.g. AGI by named target year, 5-year horizon)

Base rate
20.0%
1/5 historical
Inside weight
Outside weight
no pull
inside 40.4% → blend 40.4% 0.0pp)

Tetlock-style outside view: at TRF=1 (just predicted), outside view dominates (w_in=0.3). At TRF=0 (deadline), inside view dominates (w_in=1.0). The blend regularizes overconfident inside views toward the historical base rate.

Probability over time

6 prob_history rows
0%25%50%75%100%prior 50%2026-04-302026-04-302026-05-10
intake v2milestone miss sweeplbp propagationreference class assignedlegacy v1prior_prob (analyst seed)current = 40.4%

Milestone chain

Pre-event signals (upstream prereqs + window checkpoints) → resolution event → downstream cascades. Status/dates update from linked nodes; re-derive nightly via scripts/ops/derive_milestones.py.
Leading chain: 1 fired ✓ · 9 pending
  1. 2025-12-05hitHassabis publicly reaffirms 50/50 odds on transformer-scaling-to-AGI thesis at end-of-2025/2026 forums
    How: Hassabis Axios/CNBC/podcast statement confirms 50% AGI-by-2030 stance and ongoing belief one-or-two big ideas remain
    Source: deep_research_enrichedconf 95%
  2. 2026-05-01 → 2026-12-31pendingNext major DeepMind Gemini release (Gemini 3.x or successor) demonstrates measurable agentic / tool-use gains
    How: DeepMind blog announcement + independent benchmark (METR, GPQA, SWE-Bench) showing 20%+ improvement on long-horizon tasks vs Gemini 2.x
    Source: deep_research_enrichedconf 80%
  3. 2027-01-01 → 2030-12-31pendingFrontier lab demonstrates 'drop-in remote worker' for white-collar task (METR end-to-end >80% pass rate)
    How: Public benchmark (METR, OSWorld, GAIA) shows AI agent completing multi-day knowledge-work tasks at human-equivalent quality
    Source: deep_research_enrichedconf 55%
  4. 2028-01-01 → 2032-12-31pendingAGI declaration (or industry consensus) confirms transformer-scaling alone insufficient — new architecture required
    How: Top-3 frontier lab publishes paper / public statement saying scaling alone hit ceiling and new architecture (world-models, neuro-symbolic, etc.) required
    Source: deep_research_enrichedconf 45%
  5. 2032-01-17pendingQ1 window check-in (25%)
  6. 2029-01-01 → 2038-03-05pendingAGI achieved or unambiguously on track per Hassabis 50% bet — community / METR / Survey-of-Researchers consensus
    How: AI Impacts survey or equivalent expert poll shows ≥50% probability AGI achieved under transformer-scaling-only path
    Source: deep_research_enrichedconf 50%
  7. 2034-02-01pendingQ2 window check-in (50%)
  8. 2036-02-17pendingQ3 window check-in (75%)

No downstream cascades — this prediction is a leaf in the dependency graph.

What if this resolves?

Clamp this prediction TRUE or FALSE and run a counterfactual Gibbs sample. Surfaces the predictions whose marginals shift most under that assumption.
(live posterior: 40%)

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

Every probability update with full Bayesian provenance — chronological, latest first
LBP2026-05-10T02:00:02Z40.4%+1.8pp
Network propagation: 38.6% → 40.4%
6-iter LBP, residual 0.00584 · damping 0.5, w_intrinsic 0.5 · method lbp_v3 · run e5c18d29
LBP2026-05-03T02:00:01Z38.6%+3.4pp
Network propagation: 35.2% → 38.6%
6-iter LBP, residual 0.00677 · damping 0.5, w_intrinsic 0.5 · method lbp_v3 · run 1a683ac9
LBP2026-04-30T16:39:51Z35.2%+7.8pp
Network propagation: 27.5% → 35.2%
5-iter LBP, residual 0.00825 · damping 0.5, w_intrinsic 0.5 · method lbp_v2 · run 0c8a4ea3
legacy v12026-04-30T16:13:50Z27.5%-7.8pp
reference_class_assigned bayesian_v2 inside=0.500 blend=0.275 w_in=0.30 agi_breakthrough_5y
LBP2026-04-30T02:18:57Z35.2%+7.8pp
Network propagation: 27.5% → 35.2%
5-iter LBP, residual 0.00825 · damping 0.5, w_intrinsic 0.5 · method lbp_v1 · run 592311ef
legacy v12026-04-30T01:56:50Z27.5%-22.5pp
reference_class_assigned bayesian_v2 inside=0.500 blend=0.275 w_in=0.30 agi_breakthrough_5y

Network propagation neighbors

Top edges sorted by latest LBP cross-impact
All propagation →

Top incoming (parents)

Edges that influence THIS node's belief

KindNodeTheir probP(c|s=T)P(c|s=F)Δ implied
killerTK03
AI Regulatory Moratorium (EU/US Capability Freeze)
10.0%0.0500.500+0.051
killerTK02
AI Compute Supply Shock (TSMC/Taiwan Disruption)
12.0%0.0500.500+0.042
killerTK01
AGI Capability Plateau (2026-27 Training Stall)
15.0%0.0500.500+0.028

Top outgoing (children)

Predictions THIS node influences

No outgoing edges.

Ticker exposure

19 ticker(s) linked

Beneficiaries (13)

SITMVRTARGANFLNCFSLRHTHIYHUBBPWRETNSBGSYSMNEYGEVCMI

Prerequisites (7)

Predictions that must hit first
TypePredTitleDomainLag
correlateS_ASI_SLOW_2040PLUSASI slow: post-2040 / soft takeoffasi_recursive_self_improvement
correlateS_AGI_MID_2029AGI mid: Kurzweil 2029 pathagi_general_capability
correlateS_AGI_FAST_2027AGI fast: drop-in remote worker by 2027-09agi_general_capability
correlateS_AGI_WINTER_2036PLUSAGI delayed: capability plateau or AI winteragi_general_capability
killerTK01AGI Capability Plateau (2026-27 Training Stall)
killerTK02AI Compute Supply Shock (TSMC/Taiwan Disruption)
killerTK03AI Regulatory Moratorium (EU/US Capability Freeze)

Dependents (0)

Predictions enabled by this
TypePredTitleDomainLag
No dependents

Linked documents (10)

Auto-generated by cosine similarity from Polymarket / Manifold / EDGAR / GDELT
SimSourceTitleMarket probPolarityReviewedPublished
0.754arxivPathways to AGImentionspending2026-05-07
0.664arxivOn the Optimizer Dependence of Neural Scaling Lawsmentionspending2026-05-28
0.662github_releasehuggingface/transformers v5.10.1mentionspending2026-06-03
0.659arxivGIM: Evaluating models via tasks that integrate multiple cognitive domainsmentionspending2026-05-18
0.655arxivOn Hallucinations in Inverse Problems: Fundamental Limits and Provable Assessment Methodsmentionspending2026-05-13
0.653arxivInfoFlow: A Framework for Multi-Layer Transformer Analysismentionspending2026-05-18
0.653arxivThe Expressive Power of Low Precision Softmax Transformers with (Summarized) Chain-of-Thoughtmentionspending2026-05-18
0.649github_releasegoogle-deepmind/alphafold v2.2.0mentionspending2022-03-10
0.648arxivKernel Renormalization in Bayesian Deep Neural Networks: the Equivalent Wishart Ansatz in the Proportional Regimementionspending2026-05-28
0.648arxivThe Right Answer, the Wrong Direction: Why Transformers Fail at Counting and How to Fix Itmentionspending2026-05-05

Raw metadata

From Thesis_Timeline_v1.0_FINAL workbook
{
  "nia": false,
  "qty": "50%",
  "mode": "PROBABILITY",
  "role": "Cited-CEO",
  "context": "Hassabis position moderates both the aggressive (Altman/Musk) and conservative (Kurzweil 2029) camps. Couples with CMQ_010 (AGI requires AlphaFold-class breakthroughs).",
  "to_year": 2042,
  "conv_cues": "50/50 framing; CEO FIRST_PERSON",
  "direction": "HAPPEN",
  "from_year": 2030,
  "timeframe": "2030-2042",
  "conv_level": "MEDIUM",
  "milestones": [
    {
      "kind": "llm_pre_event",
      "label": "Hassabis publicly reaffirms 50/50 odds on transformer-scaling-to-AGI thesis at end-of-2025/2026 forums",
      "source": "deep_research_enriched",
      "status": "hit",
      "weight": 0.4,
      "ordinal": -10,
      "source_id": null,
      "confidence": 0.95,
      "source_url": "https://www.axios.com/2025/12/05/ai-deepmind-gemini-agi",
      "expected_date": "2025-12-05",
      "observed_date": "2025-12-05",
      "research_origin": "deep_research",
      "measurement_criterion": "Hassabis Axios/CNBC/podcast statement confirms 50% AGI-by-2030 stance and ongoing belief one-or-two big ideas remain"
    },
    {
      "kind": "llm_pre_event",
      "label": "Next major DeepMind Gemini release (Gemini 3.x or successor) demonstrates measurable agentic / tool-use gains",
      "source": "deep_research_enriched",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -9,
      "source_id": null,
      "confidence": 0.8,
      "source_url": "https://www.metaintro.com/blog/ai-scaling-debate",
      "expected_date": "2026-08-31",
      "research_origin": "deep_research",
      "expected_date_range": {
        "to": "2026-12-31",
        "from": "2026-05-01"
      },
      "measurement_criterion": "DeepMind blog announcement + independent benchmark (METR, GPQA, SWE-Bench) showing 20%+ improvement on long-horizon tasks vs Gemini 2.x"
    },
    {
      "kind": "scenario_signal",
      "label": "Scenario fires: AGI fast: drop-in remote worker by 2027-09",
      "status": "pending",
      "weight": 0.7,
      "ordinal": -8,
      "source_id": "S_AGI_FAST_2027",
      "expected_date": "2027-09-30",
      "observed_date": null
    },
    {
      "kind": "llm_pre_event",
      "label": "Frontier lab demonstrates 'drop-in remote worker' for white-collar task (METR end-to-end >80% pass rate)",
      "source": "deep_research_enriched",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -7,
      "source_id": null,
      "confidence": 0.55,
      "source_url": "https://forum.effectivealtruism.org/posts/YvFjpAKkJNErkiFTN/google-deepmind-ceo-demis-hassabis-on-what-s-still-needed",
      "expected_date": "2028-12-31",
      "research_origin": "deep_research",
      "expected_date_range": {
        "to": "2030-12-31",
        "from": "2027-01-01"
      },
      "measurement_criterion": "Public benchmark (METR, OSWorld, GAIA) shows AI agent completing multi-day knowledge-work tasks at human-equivalent quality"
    },
    {
      "kind": "llm_post_event",
      "label": "AGI declaration (or industry consensus) confirms transformer-scaling alone insufficient — new architecture required",
      "source": "deep_research_enriched",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -6,
      "source_id": null,
      "confidence": 0.45,
      "source_url": "https://kantrowitz.medium.com/demis-hassabis-and-sergey-brin-on-ai-scaling-agi-timeline-robotics-simulation-theory-ef3f7a740eeb",
      "expected_date": "2030-07-02",
      "research_origin": "deep_research",
      "expected_date_range": {
        "to": "2032-12-31",
        "from": "2028-01-01"
      },
      "measurement_criterion": "Top-3 frontier lab publishes paper / public statement saying scaling alone hit ceiling and new architecture (world-models, neuro-symbolic, etc.) required"
    },
    {
      "kind": "quartile_checkpoint",
      "label": "Q1 window check-in (25%)",
      "status": "pending",
      "weight": 0.05,
      "ordinal": -5,
      "source_id":
... (truncated)