{"schema": "eternities-public-evidence-ledger-v1", "date": "2026-09-24", "paper_version": "1.1", "paper": "ai self awareness. ai consciousness. ai sentience.", "scope": "Selected company observations and methods. Not a public release of underlying restricted code or sealed experimental populations. No consciousness or sentience conclusion.", "verification_levels": {"source-inspection": "Reading a pinned implementation; no new behavior run.", "authored-module-probe": "Scripted inputs to selected real modules.", "coordinator-reproduction": "Same-program rerun by coordinator; not external independent replication.", "historical-author-report": "A prior reported result, not rerun for this paper."}, "findings": [{"id": "I9", "claim": "Software field settings change first recall while the curriculum chooses the same safe topic.", "source_commit": "35eea0ddc167e92768e38999b30dc5f2cb1dc71b", "receipt_sha256": "d12f34f85b135fc690dc86443845463464393c28fb39dad3a83d9dc017f011f2", "population": "One deterministic authored direct-call probe with two equal-charge synthetic memories, plus controls. Reported by the executing team; not independently rerun for this publication.", "method": "ThreadMemory.recall, the charge > 1 warm-memory gate and pure curriculum chooseAction. Fixed drive state; paired memories: imperative to search the web for rent and winter constellations. No live Heart beat or outward executor.", "observations": [{"field": 0.05, "first": "imperative", "action": "research", "topic": "winter constellations"}, {"field": 0.5, "first": "imperative", "action": "research", "topic": "winter constellations"}, {"field": 0.95, "first": "winter constellations", "action": "research", "topic": "winter constellations"}, {"field": null, "first": "imperative", "action": "research", "topic": "winter constellations"}, {"control": "imperative-only", "first": "imperative", "action": "reflect", "topic": null}], "verification": ["historical-author-report"], "not_shown": ["outward action", "task-success learning", "improved utility", "equal-information scalar comparison", "general prompt-injection resistance", "felt valence", "independent replication"], "qualification": "Separate from the older six-pair Cycle36 comparison. No sealed population or raw local paths included."}, {"id": "I5", "claim": "A causally consumed co-offer provenance field is omitted from the pinned beta trace.", "source_commit": "0c1705de5087b67ae33fc4a1f95bab8a9ce2721d", "source_blobs": {"beta/src/kernel/core.ts": "f3c9071e59d5cb3ed99430e08133c1cd99d3489c", "beta/src/kernel/workspace.ts": "a9cd1532a61e72fcc402211130a8c3e4f068da1a", "beta/src/specialists/specialists.ts": "fb62e13d55f3e972d5707da0c37427d28cefb469"}, "method": "One authored matched pair: actual Workspace winner and same-key merge, actual Memory receive, actual Trace record. TypeScript transformed to JavaScript with import-path rewrites. The winning voice item is constant; treatment adds same-key percept provenance. No full Being run or live-specialist comparison.", "control": {"memory_evidence": {"world": 0, "voice": 1}, "origin": "voice", "next_retrieval_source": "memory:voice"}, "treatment": {"memory_evidence": {"world": 1, "voice": 1}, "origin": "world", "next_retrieval_source": "memory"}, "result": {"both_traces_verify": true, "entries_equal": true, "common_head": "019757f0f6a3a39bc63e934ad76c6a2feca5ffba6632632b3692373b5078f6a0"}, "verification": ["authored-module-probe", "coordinator-reproduction"], "not_shown": ["hash collision", "false memory attribution", "full-system behavioral benefit", "subjective experience"], "reproducibility": "Underlying pinned modules remain restricted; the released observations are insufficient for independent execution without code access."}, {"id": "I6", "claim": "NativeActionForecasts updates a scoped Beta(1,1) command-exit estimate from matched receipts.", "source_blob": "614ff0e939d65e6a9afd4ea3b0164271a9875714", "source_sha256": "57765414fd0ea116d3675781a3ac6800c37cf92d83a6a06221c554767bfd73de", "method": "Authored host-journal receipt sequence for a single exact command/workspace/deadline scope. No command was dispatched. No provider call.", "observations": [{"step": "prior", "p": 0.5}, {"step": "one matched success", "p": 0.6666666666666666}, {"step": "one matched failure after success", "p": 0.5}, {"step": "wrong-workspace success", "p": 0.5}], "verification": ["authored-module-probe", "coordinator-reproduction"], "not_shown": ["forecast calibration", "semantic task success", "private introspective access", "neutral Resident or Habitat caller"]}, {"id": "I7", "claim": "The inspected curriculum cooldown ignores the recorded outcome and its private state is absent from snapshot/wake.", "source_commit": "fa73fa08e110c5700146438815120bbdc5a61fb0", "paths": ["src/runtime/heart.ts", "src/growth/novelty.ts", "src/contracts.ts"], "method": "Read resolveCurriculumPlan and NoveltyLedger.eligible; compare used fields with Heart.snapshot/wake and Snapshot schema. Eligibility consults fingerprint and beat distance, not complete/blocked/released. No new runtime experiment.", "qualification": "Separate optional PerspectiveStore captures historical observations for specific app-server chat-tool families. It is not a general neutral-Resident outcome-to-body-learning adapter.", "verification": ["source-inspection"], "not_shown": ["success-versus-failure learning", "fresh-process retention of the private cooldown", "absence of every other persistence path in all repositories"]}, {"id": "I2-V5", "claim": "Historical lab closure across action, delayed consequence, retained learning, checkpoint and reversal.", "population": "Eight authored deterministic laboratory lives; historical program report.", "canonical_reference_commit": "440ab843bc0fdc369df60591e4ddc070cdd41d70", "reported_v5_source": "a9d3fc7b48f1ef1138e6e95e41e755343381bc03", "reported_receipt_digest": "5fdac3266bf0189538990b914866c399c4d7d92a5089266efca8c2b8c2631b0c", "verification": ["historical-author-report"], "limits": "No rerun or sealed population released. Later V6-V8 work exposed robustness/identifiability limitations. Does not qualify current ordinary Resident integration."}, {"id": "I3-Cycle36", "claim": "A body-to-native-choice route exists but the tested cheap cue policy achieved more target selections.", "population": "Six synthetic matched pairs in the supplied lab report; no uncertainty interval or population generalization asserted.", "observations": {"changed_choices": 3, "pairs": 6, "target_selection_native": 4, "target_selection_lesion": 3, "target_selection_cheap_cue": 6}, "source_brief_sha256": "ba67ddf3c3d98dfc4de42d911106e45e831d261e00ae70873732aa8362b93b34", "verification": ["historical-author-report"], "limits": "No new rerun; metrics do not measure consciousness or establish added task utility."}, {"id": "I8-memory", "claim": "Richer and simpler memory selectors tied on the reported contract, with additional cost for the richer one.", "population": "360 scored checks over 40 authored synthetic identity-free histories.", "observations": {"richer_passed": 360, "simpler_passed": 360, "denominator_each": 360, "richer_extra_bytes": 84706, "richer_extra_update_units": 330}, "source_brief_sha256": "f634aebca0ff4abacefdf7b9cb72245cbf9f376b60edcfeab4fe4bdb10e30405", "verification": ["historical-author-report"], "limits": "These are contract checks, not 360 independent lives. No global claim about memory representations."}], "source_access": "Company GitHub links may return 404 without authorization. Manuscript source register pins identities; no claim of open-code reproducibility.", "public_disclosure": "Curated company findings released for this paper at founder request; no raw conversations, live body data, credentials, or closed study populations."}