{ "schema_version": 1, "builder_plan": { "scenario": "A workflow needs specialist review, but extra agents may add more cost than useful work.", "change": "Test the proposed division of work against a single-agent implementation.", "baseline": "One agent with the same tools and an equal total token and time budget.", "steps": [ "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.", "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.", "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.", "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test." ], "measure": "Accepted outputs, coordination failures, wall-clock time, and total cost across all agents.", "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation." }, "builder_review": null, "arxiv_id": "2610.12176", "title": "Recursive Self-Improvement through Multi-Agent Self-Supervision", "summary": "MASS replaces a single model critiquing its own outputs (and hand-designed agent topologies) with an evolutionary search over multi-agent workflows, whose self-generated traces are then used for supervised fine-tuning of the base model. The cost is substantial: multiple full MASS cycles of workflow proposal, execution, self-evaluation, and SFT on a 27B model, which exceeds a single 24 GB GPU and a small cloud budget. The main failure modes are evaluator drift (the model grading itself can reward fluent but wrong reasoning on non-verifiable tasks), overfitting the evolutionary search to benchmark-like tasks, and orchestration gains that may not transfer to other models or domains.", "rationale": "The full RSI loop requires multi-round evolutionary search plus fine-tuning of a 27B model, which is infeasible on one 24 GB GPU, though the underlying idea of searching multi-agent topologies can be tracked for later lightweight adoption via APIs.", "technique": "Multi-agent self-supervision for recursive self-improvement", "family": "orquestracao_multiagente", "recommendation": "observar", "published": "2026-10-08", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.12176", "independent_implementations": 1, "total_implementations": 2, "stars": 132, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "MASS replaces a single model critiquing its own outputs (and hand-designed agent topologies) with an evolutionary search over multi-agent workflows, whose self-generated traces are then used for supervised fine-tuning of the base model. The cost is substantial: multiple full MASS cycles of workflow proposal, execution, self-evaluation, and SFT on a 27B model, which exceeds a single 24 GB GPU and a small cloud budget. The main failure modes are evaluator drift (the model grading itself can reward fluent but wrong reasoning on non-verifiable tasks), overfitting the evolutionary search to benchmark-like tasks, and orchestration gains that may not transfer to other models or domains.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.12176", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "1.2-1.6x higher performance per output token on four open-ended benchmarks; multi-agent-trained student beats single-agent student given 1.4x more training tokens.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.12176", "source_page": null, "source_excerpt": "" } ], "exposure_map": [ { "dimension": "quality", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compute", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "latency", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "operations", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compatibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "security", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "data_and_training", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "reproducibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" } ], "risks": [], "minimum_test": [], "open_questions": [], "independent_tests": [], "lab_reproductions": [], "report_available": false, "validation_tier": "unknown", "evidence_tier": "unknown", "infrastructure_basis": "unknown", "equations": [], "equations_status": "not_fetched", "core_kind": "", "equations_fetched_at": "", "equations_source_url": "", "equations_source_sha256": "", "exposures_reviewed_at": "", "exposures_source_sha256": "" }