Open Quantum Conversation
Open mode: the prompt is stored only. It is never classified, matched, weighted, or sent to Outshift. Each sentence independently starts from the entire semantic universe.
Ready.
Latest response
No draw yet.
Score before interpreting
Drive Mode
Speak your question, then say Ready. V12 answers aloud: “Ready. Waiting for the signal.” A separate Outshift timing channel determines when the automatic language draw begins. You can still say Draw for an immediate manual draw.
Voice commands: Ready · Draw · Cancel · Repeat response · Repeat question · Clear question · Stop drive mode. Keep the browser foregrounded and do not handle the phone while a vehicle is moving.
Structured Verification
Verification mode: the label/question is logged only. It does not alter selection. Each field below is independently selected from a frozen table by Outshift. Use this before naming a candidate whenever possible.
Ready.
No verification draw yet.
Neutral Name Speller
No name bank exists. Length and every A–Z letter are independently selected by Outshift. No autocomplete, dictionary, phonetic matching, rearranging, or fuzzy rescue.
No spelling draw yet.
Session Log
Session Analysis
Recurrence statistics describe the configured random null. They are not probabilities that a spirit, consciousness, or identity is present.
Macro families
Semantic domains
Communicative intents
Valence-bearing details
Signal windows
Semantic Map
Open conversation is hierarchical and prompt-independent: 1/12 speech family → 1/8 speech act (96 acts total), then for topic-bearing acts 1/8 macro → 1/8 domain → 1/32 subdomain (2,048 literal topic leaves) → 1/8 topic focus. Expression form is selected separately from 16 audited surface forms. The renderer may reorder or phrase selected meaning, but it may not invent semantic claims.
Speech families & acts
Macro families, domains & subdomains
Structured verification banks
Ontology audit
Version: V12.0.0-EXPANDED-VERIFY
Ontology SHA-256: 35a49044e313b9e62633a1cb3fdcf565ce14595e0a1a653e5e83ccaa8bbb3ddc
Macro families: 8
Domains: 64
Open-conversation leaf topics: 1024
Structured packet banks: IDENTITY · REALITY · EVENT · HARD
Prompt matching: NONE
Prompt classification: NONE
Linked sentences: NONE
Topic persistence: NONE
Named-person bank: NONE
Experimental RNG: Cisco Outshift QRNG only
Settings
Not tested.
Timing checks occur every 4 seconds. Standard trigger rule: one 8-bit Outshift value in 0–15 (1/16) opens the signal window. Controls use a precomputed matched geometric wait from a separate Outshift stream.
Protocol / audit rules
1. V12 is a new frozen ontology. V11 remains unchanged and uses its own storage key and ontology hash.
2. User prompt text is logged only. It is not classified, embedded, matched, weighted, or sent to Outshift.
3. Open Conversation first selects one of 12 speech families, then one of 8 speech acts inside that family: 96 equally reachable speech acts.
4. Topic-bearing acts independently select 1 of 8 macro families, 1 of 8 domains, and 1 of 32 subdomains: 2,048 literal topic leaves.
5. A separate 1-of-8 TOPIC FOCUS selects whether the sentence concerns the core subject, person, action, timing, place, cause, outcome, or relationship.
6. A separate 1-of-16 EXPRESSION FORM determines surface phrasing/order. Expression is logged and cannot add facts absent from the semantic packet.
7. The old blanket phrase “I do not think X is the right interpretation” is removed. Corrections now specify what is being rejected: topic, person, event, cause, time, place, or relationship.
8. Meaning is selected before rendering. Semantic packets are authoritative; visible English is only the deterministic realization of selected fields.
9. Conversation banks are expanded for ordinary life, identity, family, caregiving, work, medicine, disability, transportation, media, religion, death, other realities, nonhuman hypotheses, and null/non-spiritual explanations.
10. Structured Verify now includes IDENTITY, REALITY, EVENT, BODY/HEALTH, RELATIONSHIP, DEATH, COMMUNICATION, HARD VALUES, and DIRECT ANSWER modes.
11. Metaphysical options are hypotheses, not premises. Null options include coincidence, pattern matching, memory reconstruction, software/renderer errors, QRNG/API irregularity, and no external communicator.
12. Direct-answer values render explicitly. Unknown values fail closed; undefined output is not accepted.
13. Abstract language limits are explicit fields rather than malformed sentence fragments.
14. Adult sexuality remains non-graphic. Childhood sexual harm remains a non-graphic abuse category only.
15. Earth is not assumed. Alternate Earth, non-Earth physical world, other physical reality, nonphysical setting, symbolic setting, and no-external-reality options remain available.
16. Drive speech cleanup merges overlapping recognition revisions only. It does not paraphrase or semantically rewrite speech.
17. READY timing remains logically separate from language selection. Matched-control timing remains available.
18. Outshift failure fails closed. There is no fallback language RNG.
19. Packet-path odds describe the exact configured selector path only. They are not coherence p-values, identity probabilities, or proof of an external source.
20. Do not discard misses, contradictions, null outputs, or hard-value failures. They are part of the experiment.
21. V12 was designed after all earlier sessions. Newly added vocabulary cannot be retroactively counted as evidence in older versions.