{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:DZ45ASKHJJZWLYBZWTRBQEWVI4","short_pith_number":"pith:DZ45ASKH","canonical_record":{"source":{"id":"2505.22954","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-29T00:26:15Z","cross_cats_sorted":[],"title_canon_sha256":"6fc7af6007f2e0ba1d72d6948e35470cf38dd23ed22526ba2bd27e572c17b907","abstract_canon_sha256":"d608354ef48d0920568eecd0242f28c2130580f0a0ae75d15949fe6db01645db"},"schema_version":"1.0"},"canonical_sha256":"1e79d049474a7365e039b4e21812d54704eaa66687e0bc02ffb1df904ced2bae","source":{"kind":"arxiv","id":"2505.22954","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.22954","created_at":"2026-05-17T23:38:49Z"},{"alias_kind":"arxiv_version","alias_value":"2505.22954v3","created_at":"2026-05-17T23:38:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.22954","created_at":"2026-05-17T23:38:49Z"},{"alias_kind":"pith_short_12","alias_value":"DZ45ASKHJJZW","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"DZ45ASKHJJZWLYBZ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"DZ45ASKH","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:DZ45ASKHJJZWLYBZWTRBQEWVI4","target":"record","payload":{"canonical_record":{"source":{"id":"2505.22954","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-29T00:26:15Z","cross_cats_sorted":[],"title_canon_sha256":"6fc7af6007f2e0ba1d72d6948e35470cf38dd23ed22526ba2bd27e572c17b907","abstract_canon_sha256":"d608354ef48d0920568eecd0242f28c2130580f0a0ae75d15949fe6db01645db"},"schema_version":"1.0"},"canonical_sha256":"1e79d049474a7365e039b4e21812d54704eaa66687e0bc02ffb1df904ced2bae","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:49.077031Z","signature_b64":"MZicThfsgCEA5pIsYAcjsxdzl28UL4hYwvK4cDIJSHdHQNNXIJ/SIUg50UiZrDEkBMP8vyot5l7tUvwYNDHuBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1e79d049474a7365e039b4e21812d54704eaa66687e0bc02ffb1df904ced2bae","last_reissued_at":"2026-05-17T23:38:49.076403Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:49.076403Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.22954","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:49Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ALG13xwjch/VOEScsMAKQ4iQUETSekmZtb3239aPaE+IjYc68KlAjZ2ykMONoMNw5Kxv6YIOdz7r6CYI6+PAAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T22:10:44.595247Z"},"content_sha256":"e53e43699384dab7567971573d95c3629e78ad5b262c51b0505f491358e9ad80","schema_version":"1.0","event_id":"sha256:e53e43699384dab7567971573d95c3629e78ad5b262c51b0505f491358e9ad80"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:DZ45ASKHJJZWLYBZWTRBQEWVI4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"The Darwin Godel Machine lets AI agents iteratively rewrite their own code and validate each change on benchmarks, raising SWE-bench performance from 20 percent to 50 percent.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Cong Lu, Jeff Clune, Jenny Zhang, Robert Lange, Shengran Hu","submitted_at":"2025-05-29T00:26:15Z","abstract_excerpt":"Today's AI systems have human-designed, fixed architectures and cannot autonomously and continuously improve themselves. The advance of AI could itself be automated. If done safely, that would accelerate AI development and allow us to reap its benefits much sooner. Meta-learning can automate the discovery of novel algorithms, but is limited by first-order improvements and the human design of a suitable search space. The G\\\"odel machine proposed a theoretical alternative: a self-improving AI that repeatedly modifies itself in a provably beneficial manner. Unfortunately, proving that most change"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Empirically, the DGM automatically improves its coding capabilities (e.g., better code editing tools, long-context window management, peer-review mechanisms), increasing performance on SWE-bench from 20.0% to 50.0%, and on Polyglot from 14.2% to 30.7%.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That repeated empirical validation on the chosen benchmarks is sufficient to guarantee net-beneficial self-modifications without formal proof or exhaustive testing of downstream effects.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"The Darwin Gödel Machine evolves its own code via open-ended archive-based exploration with foundation models, raising SWE-bench performance from 20.0% to 50.0% and Polyglot from 14.2% to 30.7%.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"The Darwin Godel Machine lets AI agents iteratively rewrite their own code and validate each change on benchmarks, raising SWE-bench performance from 20 percent to 50 percent.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"6391325b72921065c028c57311a970746d2545d19011936bf0ddf39f05d3019d"},"source":{"id":"2505.22954","kind":"arxiv","version":3},"verdict":{"id":"dae3ea44-8bb4-475f-8dec-57aacc7a19f8","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T04:36:01.490864Z","strongest_claim":"Empirically, the DGM automatically improves its coding capabilities (e.g., better code editing tools, long-context window management, peer-review mechanisms), increasing performance on SWE-bench from 20.0% to 50.0%, and on Polyglot from 14.2% to 30.7%.","one_line_summary":"The Darwin Gödel Machine evolves its own code via open-ended archive-based exploration with foundation models, raising SWE-bench performance from 20.0% to 50.0% and Polyglot from 14.2% to 30.7%.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That repeated empirical validation on the chosen benchmarks is sufficient to guarantee net-beneficial self-modifications without formal proof or exhaustive testing of downstream effects.","pith_extraction_headline":"The Darwin Godel Machine lets AI agents iteratively rewrite their own code and validate each change on benchmarks, raising SWE-bench performance from 20 percent to 50 percent."},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":3,"snapshot_sha256":"cd3fc07a273f512a1936ad0b5c040133ec6547ab9d0db223063903c84bb8c0ed"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"verdict_id":"dae3ea44-8bb4-475f-8dec-57aacc7a19f8"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:38:49Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7mJF5D7nOKvKial/xZ5sxEhudKf8rmXpvJMfK4NO//mny+K9tF6rngbaR9BlE1Xy7g9Ss2jwcSboK4xaKWSACA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-03T22:10:44.595743Z"},"content_sha256":"36e60b8a9163249464a52caf0b9273fd0ed65d1bcd4a9fede5ea50ad117c37ce","schema_version":"1.0","event_id":"sha256:36e60b8a9163249464a52caf0b9273fd0ed65d1bcd4a9fede5ea50ad117c37ce"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DZ45ASKHJJZWLYBZWTRBQEWVI4/bundle.json","state_url":"https://pith.science/pith/DZ45ASKHJJZWLYBZWTRBQEWVI4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DZ45ASKHJJZWLYBZWTRBQEWVI4/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-03T22:10:44Z","links":{"resolver":"https://pith.science/pith/DZ45ASKHJJZWLYBZWTRBQEWVI4","bundle":"https://pith.science/pith/DZ45ASKHJJZWLYBZWTRBQEWVI4/bundle.json","state":"https://pith.science/pith/DZ45ASKHJJZWLYBZWTRBQEWVI4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DZ45ASKHJJZWLYBZWTRBQEWVI4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:DZ45ASKHJJZWLYBZWTRBQEWVI4","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"d608354ef48d0920568eecd0242f28c2130580f0a0ae75d15949fe6db01645db","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-29T00:26:15Z","title_canon_sha256":"6fc7af6007f2e0ba1d72d6948e35470cf38dd23ed22526ba2bd27e572c17b907"},"schema_version":"1.0","source":{"id":"2505.22954","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.22954","created_at":"2026-05-17T23:38:49Z"},{"alias_kind":"arxiv_version","alias_value":"2505.22954v3","created_at":"2026-05-17T23:38:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.22954","created_at":"2026-05-17T23:38:49Z"},{"alias_kind":"pith_short_12","alias_value":"DZ45ASKHJJZW","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"DZ45ASKHJJZWLYBZ","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"DZ45ASKH","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:36e60b8a9163249464a52caf0b9273fd0ed65d1bcd4a9fede5ea50ad117c37ce","target":"graph","created_at":"2026-05-17T23:38:49Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Empirically, the DGM automatically improves its coding capabilities (e.g., better code editing tools, long-context window management, peer-review mechanisms), increasing performance on SWE-bench from 20.0% to 50.0%, and on Polyglot from 14.2% to 30.7%."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That repeated empirical validation on the chosen benchmarks is sufficient to guarantee net-beneficial self-modifications without formal proof or exhaustive testing of downstream effects."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"The Darwin Gödel Machine evolves its own code via open-ended archive-based exploration with foundation models, raising SWE-bench performance from 20.0% to 50.0% and Polyglot from 14.2% to 30.7%."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"The Darwin Godel Machine lets AI agents iteratively rewrite their own code and validate each change on benchmarks, raising SWE-bench performance from 20 percent to 50 percent."}],"snapshot_sha256":"6391325b72921065c028c57311a970746d2545d19011936bf0ddf39f05d3019d"},"formal_canon":{"evidence_count":3,"snapshot_sha256":"cd3fc07a273f512a1936ad0b5c040133ec6547ab9d0db223063903c84bb8c0ed"},"paper":{"abstract_excerpt":"Today's AI systems have human-designed, fixed architectures and cannot autonomously and continuously improve themselves. The advance of AI could itself be automated. If done safely, that would accelerate AI development and allow us to reap its benefits much sooner. Meta-learning can automate the discovery of novel algorithms, but is limited by first-order improvements and the human design of a suitable search space. The G\\\"odel machine proposed a theoretical alternative: a self-improving AI that repeatedly modifies itself in a provably beneficial manner. Unfortunately, proving that most change","authors_text":"Cong Lu, Jeff Clune, Jenny Zhang, Robert Lange, Shengran Hu","cross_cats":[],"headline":"The Darwin Godel Machine lets AI agents iteratively rewrite their own code and validate each change on benchmarks, raising SWE-bench performance from 20 percent to 50 percent.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-29T00:26:15Z","title":"Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.22954","kind":"arxiv","version":3},"verdict":{"created_at":"2026-05-16T04:36:01.490864Z","id":"dae3ea44-8bb4-475f-8dec-57aacc7a19f8","model_set":{"reader":"grok-4.3"},"one_line_summary":"The Darwin Gödel Machine evolves its own code via open-ended archive-based exploration with foundation models, raising SWE-bench performance from 20.0% to 50.0% and Polyglot from 14.2% to 30.7%.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"The Darwin Godel Machine lets AI agents iteratively rewrite their own code and validate each change on benchmarks, raising SWE-bench performance from 20 percent to 50 percent.","strongest_claim":"Empirically, the DGM automatically improves its coding capabilities (e.g., better code editing tools, long-context window management, peer-review mechanisms), increasing performance on SWE-bench from 20.0% to 50.0%, and on Polyglot from 14.2% to 30.7%.","weakest_assumption":"That repeated empirical validation on the chosen benchmarks is sufficient to guarantee net-beneficial self-modifications without formal proof or exhaustive testing of downstream effects."}},"verdict_id":"dae3ea44-8bb4-475f-8dec-57aacc7a19f8"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:e53e43699384dab7567971573d95c3629e78ad5b262c51b0505f491358e9ad80","target":"record","created_at":"2026-05-17T23:38:49Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"d608354ef48d0920568eecd0242f28c2130580f0a0ae75d15949fe6db01645db","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-29T00:26:15Z","title_canon_sha256":"6fc7af6007f2e0ba1d72d6948e35470cf38dd23ed22526ba2bd27e572c17b907"},"schema_version":"1.0","source":{"id":"2505.22954","kind":"arxiv","version":3}},"canonical_sha256":"1e79d049474a7365e039b4e21812d54704eaa66687e0bc02ffb1df904ced2bae","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1e79d049474a7365e039b4e21812d54704eaa66687e0bc02ffb1df904ced2bae","first_computed_at":"2026-05-17T23:38:49.076403Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:38:49.076403Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"MZicThfsgCEA5pIsYAcjsxdzl28UL4hYwvK4cDIJSHdHQNNXIJ/SIUg50UiZrDEkBMP8vyot5l7tUvwYNDHuBg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:38:49.077031Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.22954","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e53e43699384dab7567971573d95c3629e78ad5b262c51b0505f491358e9ad80","sha256:36e60b8a9163249464a52caf0b9273fd0ed65d1bcd4a9fede5ea50ad117c37ce"],"state_sha256":"a26382f5faf99d62c3779d944f8821e55153df9586d735fd64cd80ee477afdea"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qQrlpCLeq7u2yNNL3zNv+0DITnGcmaebk+AZfBgBBztPv9lE8FiDAh7t+y0MARNywEzJXquDpvcvBHzN5kr4Dg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-03T22:10:44.598037Z","bundle_sha256":"494a01efa0ca44fdaf0d6d2223583d6fb54bb8f38e28269c5d7dab7bdf87e77d"}}