{
  "schema": "vedokrok.public-item.v1",
  "release_id": "MHC-RPUB-20260920-75ad787a",
  "url": "/knowledge/parallelize-agents-only-across-separable-work",
  "id": "MHC-D-RESEARCH-0716",
  "version": "0.1.0",
  "title": "Parallelize agents only across separable work",
  "summary": "Sixteen agents cannot parallelize one locked door.",
  "kind": "heuristic",
  "body": "Find independent tests, modules, research questions or artifacts that can progress without repeatedly editing the same dependency. Assign those in parallel. Keep serial bottlenecks serial until you can split them with an external oracle or clearer interface. More workers on one tightly coupled failure often create duplicate effort and conflicting edits.",
  "limits": [
    "Parallelism can increase cost faster than throughput; compare elapsed-time savings with merge, review and compute overhead."
  ],
  "topics": [
    "union-ai-coding-harness-work"
  ],
  "intents": [],
  "source_ids": [
    "RS-7CFA09733D80048C"
  ],
  "evidence": [
    {
      "claim": "Anthropic's parallel-agent compiler experiment found parallelism useful when agents could work on distinct failing tests or specialized tasks, but wasteful when all agents hit the same serial bottleneck.",
      "source_id": "RS-7CFA09733D80048C",
      "role": "supports",
      "note": "The experiment is unusually large and expensive; ordinary projects may not justify many parallel agents.",
      "locator": "Make parallelism easy; Linux-kernel bottleneck"
    }
  ],
  "use_when": [
    "You want multiple agents to accelerate a large implementation or investigation."
  ],
  "avoid_when": [
    "Parallelism can increase cost faster than throughput; compare elapsed-time savings with merge, review and compute overhead."
  ],
  "example": "Agents can fix independent failing tests in parallel; they should not all rewrite the same shared parser at once without partitioning or interface constraints.",
  "check": "Parallel workers have mostly independent failure surfaces and can produce mergeable evidence.",
  "sources": [
    {
      "id": "RS-7CFA09733D80048C",
      "title": "Building a C compiler with a team of parallel Claudes",
      "url": "https://www.anthropic.com/engineering/building-c-compiler"
    }
  ],
  "relations": [
    {
      "from": "MHC-D-RESEARCH-0716",
      "to": "MHC-D-RESEARCH-0717",
      "type": "use_before",
      "url": "/knowledge/give-parallel-agents-independent-tests-and-merge-boundaries"
    }
  ],
  "collections": [
    {
      "id": "RC-A76353E5E9A33FE5",
      "title": "Make coding agents work in inspectable increments",
      "url": "/collections/make-coding-agents-work-in-inspectable-increments"
    }
  ]
}