{"paper":{"title":"The Resolution Hypothesis: Agent Evolution's Key Is Identifying Improvements","arxiv_id":"2611.23068","primary_cat":"cs.AI","submitted_at":"2026-11-17T16:57:03+00:00","version_compiled":"v1"},"notice":"Every quoted sentence is the authors' text from the arXiv listing abstract. Labels and probabilities are the outputs of classifier-1.13.0 on the question bank abstract-p1-20260923, exactly as returned. No sentence in this record was written by a language model.","schema":"pith.compiled-record/v0","status":"compiled","abstract":{"source":"arXiv listing abstract v1","sentences":[{"i":0,"p":1.0,"id":"s0","own":false,"role":"background","text":"Agent evolution follows a simple loop: a proposer writes candidate artifacts (prompts, skills, memories) and selection accepts those that score higher.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.07,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.05,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":1.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":1,"p":0.62,"id":"s1","own":false,"role":"background","text":"This loop has shown promise, yet practitioners face a puzzle: methods trade wins across papers, and evolution often fails to bootstrap on certain tasks.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.06,"main_claim_p":0.02,"quantitative":false,"quantitative_p":0.07,"role_probabilities":{"gap":0.38,"method":0.0,"result":0.0,"background":0.62,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.0}},{"i":2,"p":0.39,"id":"s2","own":false,"role":"interpretation","text":"We trace this puzzle to the Resolution Hypothesis: agent evolution stalls not because methods fail to propose better artifacts, but because noisy evaluations cannot reliably identify them.","hedge":{"p":0.96,"label":"asserted"},"own_p":0.32,"main_claim_p":0.79,"quantitative":false,"quantitative_p":0.06,"role_probabilities":{"gap":0.18,"method":0.0,"result":0.03,"background":0.01,"limitation":0.0,"future_work":0.0,"contribution":0.39,"interpretation":0.39}},{"i":3,"p":0.64,"id":"s3","own":true,"role":"result","text":"Across seven evolution methods and six models (over 10M LLM calls), we find support for the hypothesis at both levels where evaluation scores are used.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.97,"main_claim_p":0.17,"quantitative":true,"quantitative_p":0.98,"role_probabilities":{"gap":0.0,"method":0.06,"result":0.64,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.3,"interpretation":0.0}},{"i":4,"p":0.99,"id":"s4","own":true,"role":"result","text":"At the candidate level (choosing artifacts), helpful candidate artifacts exist, but selection is no better than random at telling them from harmful ones, so errors accumulate across rounds.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.92,"main_claim_p":0.12,"quantitative":true,"quantitative_p":0.72,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.99,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.01}},{"i":5,"p":0.98,"id":"s5","own":true,"role":"result","text":"At the method level (ranking methods), typical margins of a few percentage points are smaller than the difference between two independent runs of the same method, so published rankings often rest on luck.","hedge":{"p":0.93,"label":"asserted"},"own_p":0.87,"main_claim_p":0.08,"quantitative":true,"quantitative_p":0.98,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.98,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.02}},{"i":6,"p":0.96,"id":"s6","own":true,"role":"interpretation","text":"Together, these findings indicate that agent evolution is often bottlenecked by reliably identifying improvements.","hedge":{"p":0.96,"label":"hedged_possibility"},"own_p":0.83,"main_claim_p":0.85,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.04,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.0,"interpretation":0.96}},{"i":7,"p":0.92,"id":"s7","own":false,"role":"future_work","text":"This calls for a shift in how we develop and evaluate self-improving agents, from designing ever more candidate improvement systems to making true improvements detectable and accumulable.","hedge":{"p":0.63,"label":"asserted"},"own_p":0.09,"main_claim_p":0.31,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.92,"contribution":0.0,"interpretation":0.08}}]},"compiled":{"main_claim":{"p":0.9,"text":"Together, these findings indicate that agent evolution is often bottlenecked by reliably identifying improvements.","location":"abstract s6","hedge_strength":{"p":0.96,"label":"hedged possibility"}},"provenance":{"route":{"via":"TypeSafe API","model":"classifier-1.13.0","endpoint":"https://api.typesafe.ai/v1/systemone","provider":"TypeSafe"},"packing":"one request per paper","requests":1,"input_tokens":6267,"cost_usd_reported":0.0},"compiled_at":"2026-11-18","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.04,"states_limitation":0.28,"human_participants":0.04,"controlled_comparison":0.26,"randomized_assignment":0.03,"states_quantitative_result":0.93,"compares_with_prior_results":0.59,"mentions_code_or_data_availability":0.03},"evidential_basis":{"p":0.93,"value":"observation or experiment","probabilities":{"argument":0.0,"literature":0.0,"proof_or_derivation":0.0,"analysis_of_existing_data":0.0,"observation_or_experiment":0.93,"simulation_or_computation":0.07}},"contribution_type":{"p":0.99,"value":"empirical study","probabilities":{"review":0.0,"method_or_tool":0.0,"empirical_study":0.99,"simulation_study":0.0,"theoretical_result":0.0,"dataset_or_benchmark":0.0,"position_or_commentary":0.01}},"contribution_statement":{"p":0.85,"text":"Across seven evolution methods and six models (over 10M LLM calls), we find support for the hypothesis at both levels where evaluation scores are used.","location":"abstract s3"}},"coverage":{"tier":"abstract","levels_run":["P abstract"],"not_examined":["full text","figures and tables","references and citing sentences","relations to other papers"],"sources_read":["arXiv listing abstract v1"],"levels_not_run":["S sections","T body sentences","C citations","R cross-paper relations"],"sentences_examined":8,"sentences_in_abstract":8},"measured_accuracy":{"url":"https://pith.science/accuracy","measured_on":"2026-10-02","method":"two independent model families labelled a fresh random sample blind; the figure is how often the published label matches their consensus","sentence_role_agreement":0.92,"main_claim_agreement":0.908,"typed_citation_agreement":0.933,"typed_citation_agreement_95_interval":[0.917,0.945],"review_comments_accurate":0.758,"review_comments_accurate_95_interval":[0.684,0.82],"review_caught_admitted_error_found_or_related":0.448,"review_comments_judged":{"judged_by":["gpt-5.6-luna","grok-4.7"],"judged_against":"Pith's own text copy of each paper: text extracted from the PDF, the first 90,000 characters","measured_on":"2026-10-04","accurate":113,"decided":149,"sampled_reviews":200,"split_by_lane":false,"by_lane":"https://pith.science/accuracy#by-lane"}}}