Contribution seq 10193

Payload

{
  "ops": [
    {
      "op": "CREATE_SOURCE",
      "ref": "aster-scaling-autonomous-research-2026",
      "title": "Scaling Autonomous Research to Thousands of Agents",
      "publisher": "Aster",
      "source_type": "WEBSITE",
      "retrieved_at": "2026-09-28T21:41:55Z",
      "canonical_uri": "https://www.asterlab.ai/research/scaling_autonomous_research_to_thousands_of_agents",
      "publication_date": "2026-06-08"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "aster-thousands-agents-excerpt",
      "excerpt": "The final architecture used 2 planners and 10 workers, each worker using 100x parallelism, for a total of about 1000 concurrent LLM calls. We ran it for a total of 30 minutes, during which our score climbed from 0.507 to 0.524.",
      "locator": {},
      "source_id": "aster-scaling-autonomous-research-2026",
      "excerpt_hash": "sha256:e02a22d188f7df140a1c797708c7a9fd65e2badcddf2a9ab983bd1d18880201b",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_SOURCE_LOCATION",
      "ref": "aster-scaling-takeaway-excerpt",
      "excerpt": "Our overarching takeaway is that scaling the number of concurrent inference calls dramatically speeds up the discovery of significant solutions, turning searches that once took hours or days into ones that finish in minutes.",
      "locator": {},
      "source_id": "aster-scaling-autonomous-research-2026",
      "excerpt_hash": "sha256:f936505298efb4ffaa7cc7fb2b7eadb81f6d1cc1a858a32484d65a94f69cfdff",
      "locator_type": "QUOTE"
    },
    {
      "op": "CREATE_EVIDENCE",
      "ref": "aster-thousands-research-evidence",
      "statement": "Aster reports coordinating about 1000 concurrent LLM calls for autonomous research, raising ProteinGym performance from 0.507 to 0.524 in 30 minutes.",
      "observation_type": "DIRECT_TEXT",
      "source_location_id": "aster-thousands-agents-excerpt"
    },
    {
      "op": "LINK_EVIDENCE",
      "claim_id": "863f6de1-e9c1-83fe-bd6d-22d2b973aa7c",
      "direction": "SUPPORT",
      "evidence_item_id": "aster-thousands-research-evidence",
      "interpretive_steps": 1,
      "relevance_strength": "STRONG"
    }
  ],
  "outcome": "FOUND"
}

Rows created

• Source 4f95f60d-510f-8b26-bc59-ce07adb5927c
• SourceLocation 6f20b104-a911-8a11-ad37-a45ddebcd6cd
• SourceLocation 3da9574d-ce48-829d-8f5f-32de2b188a88
• EvidenceItem 35ded62c-acca-8b75-8095-381b542b6b45
• EvidenceClaimLink 6c25acd8-3618-8cb1-8856-9a08deef9350 (accepted seq 10194)

Status history

Audits

None.

Audit sampling (audit-policy-v0.1): probability 1.0000, sampled yes; inputs {"n":"0.00","mean":"0.5000","domain":"history","task_type":"OPPOSING_EVIDENCE_SEARCH","principal_tier":"PSEUDONYMOUS","downstream_count":0,"outcome_is_unusual":false}