True AGI arrives 2030-2035 (5-10 years from 2026) based on advanced 'world model systems' — current LLM architecture insufficient for generalized autonomy due to lack of temporal consistency and deep reasoning; world models inherently understand/simula...

True AGI arrives 2030-2035 (5-10 years from 2026) based on advanced 'world model systems' — current LLM architecture insufficient for generalized autonomy due to lack of temporal consistency and deep reasoning; world models inherently understand/simulate/predict complex physics of material world; within bio-sensing + molecular biology, autonomous systems will rapidly simulate protein folding, chemical interactions, climate dynamics with absolute precision, solving intractable problems before generalized superintelligence. | Next DeepMind world-model benchmark release

Key catalyst: Next DeepMind world-model benchmark release

Watch events: DeepMind world-model releases; Genie 4, Veo 3

Hassabis 2030-2035 framing consistent with 2024 Nobel interview; DeepMind Genie 3, Veo, Gemini world-modeling progress.

Adverse (4)

{
  "nia": false,
  "qty": "5-10yr AGI timeline",
  "mode": "FORECAST",
  "role": "Cited-CEO",
  "context": "Third Hassabis entry (CYB_019 world model systems, ROB_004 solve physics 2027-28, ROB_005 physical intelligence AGI prerequisite). Specific 2030-2035 AGI target + world-model-before-ASI sequencing.",
  "to_year": 2035,
  "conv_cues": "CEO FIRST_PERSON; specific year range",
  "direction": "HAPPEN",
  "from_year": 2030,
  "timeframe": "2030-2035",
  "conv_level": "HIGH",
  "milestones": [
    {
      "kind": "llm_pre_event",
      "label": "DeepMind ships next-generation Genie/Veo world-model with measurable physics-consistency benchmark improvement over Genie 3",
      "source": "https://lexfridman.com/demis-hassabis-2-transcript/",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -11,
      "source_id": null,
      "confidence": 0.7,
      "expected_date": "2027-03-17",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2027-12-31",
        "from": "2026-06-01"
      },
      "measurement_criterion": "DeepMind publishes paper or product release for a successor world-model with reported gains on a public physical-reasoning or world-simulation benchmark (e.g., RoboSuite, RT-2 successor, or new DeepMind-proposed benchmark)"
    },
    {
      "kind": "llm_pre_event",
      "label": "AlphaFold-class system released for non-protein scientific domain (materials, climate, or chemistry)",
      "source": "https://www.cst.cam.ac.uk/nobel-laureate-and-cambridge-university-alumnus-sir-demis-hassabis-heralds-new-era-ai-drug-discovery",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -10,
      "source_id": null,
      "confidence": 0.65,
      "expected_date": "2027-11-01",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2028-12-31",
        "from": "2026-09-01"
      },
      "measurement_criterion": "DeepMind announces an AlphaFold-equivalent breakthrough model for materials science, climate dynamics, or autonomous chemistry simulation with peer-reviewed accuracy claims"
    },
    {
      "kind": "llm_pre_event",
      "label": "Public long-horizon agent benchmark passes 80% on tasks that current 2025 SOTA scores under 30%",
      "source": "https://www.flowhunt.io/blog/the-decade-of-ai-agents-andrej-karpathy-agi-timeline/",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -9,
      "source_id": null,
      "confidence": 0.55,
      "expected_date": "2028-09-30",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2030-06-30",
        "from": "2027-01-01"
      },
      "measurement_criterion": "Frontier lab (DeepMind, OpenAI, Anthropic) reports >=80% on a recognized long-horizon agentic benchmark (e.g., GAIA-style, OSWorld, SWE-bench Verified-multi-day variants) where 2025 SOTA was sub-30%"
    },
    {
      "kind": "llm_pre_event",
      "label": "DeepMind / Hassabis publicly reaffirms or revises 2030-2035 AGI horizon based on world-model progress",
      "source": "https://eu.36kr.com/en/p/3598888503902981",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -8,
      "source_id": null,
      "confidence": 0.7,
      "expected_date": "2028-12-31",
      "research_origin": "training",
      "expected_date_range": {
        "to": "2030-12-31",
        "from": "2027-01-01"
      },
      "measurement_criterion": "Hassabis interview/keynote where he places explicit revised probability or year for AGI achievement, citing world-model maturity"
    },
    {
      "kind": "scenario_signal",
      "label": "Scenario fires: AGI mid: Kurzweil 2029 path",
      "status": "pending",
      "weight": 0.7,
      "ordinal": -7,
      "source_id": "S_AGI_MID_2029",
      "expected_date": "2029-03-31",
      "observed_date": null
    },
    {
      "kind": "llm_pre_event",
      "label": "AI system demonstrates novel scientific discovery autonomously (without targeted human prompting)
... (truncated)