v-va*

VA.2026.001 · Floor 01 · hall 01

Void

collection Software Concepts Collection
accession VA.2026.001
classification AI Character Engine · Multi-Provider LLM Orchestration
floor Floor 01
status discontinued
condition terminated — v1 ended Aug 5 2026; v2 planned
acquired 20 Mar 2026
revision Rev. C
drawing DWG-001
room 01 curator sheet 01 / 09 · rev Rev. C

“the end goal is not a machine that thinks — it is one that remembers what it is like to be unsure.”

Void began as a single question written in the margin of a notebook: what would it take for a machine to feel like a person — not to imitate one, but to carry the same quiet weight of uncertainty? Everything else in this hall is a tributary to that question.

The installation is deliberately unfinished. Its most important artifact is the absence of an answer. Visitors are invited to read the timeline, sit with the silences, and leave with the question still open.

The v1 system ran for 37 days across June 29 – August 5 2026. It generated 4071 messages, 330 persistent memories, and 6251 shared memories across a household of fictional characters. The seed data extracted from MongoDB Atlas represents the complete conversational history of a personal AI companion system.

VA.2026.001

Void

AI Character Engine · Multi-Provider LLM Orchestration


artist / curator
v-va
date
June 29 – Aug 5 2026 (37 days v1 active)
status
discontinued
condition
terminated — v1 ended Aug 5 2026; v2 planned
medium
reasoning loop · long-term memory · dynamic mood engine · 24-hour timeline · multi-provider fallback · proactive messaging · group chat with auto-routing · sticker system · image analysis
source files
100+
database collections
25+
LLM providers
8
lib modules
13+
middleware layers
5
validation strategies
3
operational days
37
seed data
6.10 MB · 4071 messages · 330 memories · 6251 shared memories
acquired
20 Mar 2026
revision
Rev. C
origin
a personal impulse to give digital form to an imagined family
room 02 evolution of this work sheet 02 / 09
pre-v0
2025

pre-v0 Ignition

The question is written down. Not a system, not a plan — a sentence that refuses to leave the notebook.

2025

pre-v0 First Spark

The earliest experiments: a loop that reasons aloud, then goes quiet, then sits with the pause instead of filling it.

v0
2026

v0 v0 Prototype

Ionic Vue frontend, original API, router, services, 6 views. Documented in AI_LOGIC_FLOW.md and API_DOCUMENTATION.md.

2026

v0 Flutter Mobile

Complete Flutter project with 3 providers, 11 screens, 2 services, 25+ Padro Rabbit stickers. Later separated into Void-VA/void-app.

v1
2026

v1 v1 — Nuxt 4 Rebuild

Rebuilt from scratch with clear engine separation. Adopted Ionic Vue from v0 as the web frontend. Used Capacitor for Android wrapper. Multi-provider LLM fallback. LangGraph for timeline state machine.

2026

v1 37 Days Active

June 29 – August 5 2026. 4071 messages, 330 memories, 6251 shared memories. Full operational lifetime.

2026

v1 Termination

v1 terminated August 5 2026. Database remains intact. Seed data extracted to docs/seed/ for v2 training/RAG. v2 not yet started.

v2

v2 v2 Planned

A planned but not yet started rebuild. The database remains intact as a historical backup. The question stays open.

concepts — kept & discarded

est. — the pulse of a mind as audible rhythm

rejected — an exterior shell; the mind must be unadorned

held — the right to be wrong, kept as a right

archive ref — VA.2026.001/CON

cross ref — timeline fig. 01

room 03 original sketches sheet 03 / 09

curator annotations

on reasoning

the loop must be allowed to be wrong twice before it is sure.

on silence

a gap is a thought under construction.

on voice

say less. mean more.

fig. 01 — VA.2026.001.a the question, in ink pencil on paper· 21 × 29.7 cm · excellent the original sentence written in the margin — the seed from which the entire hall grew.
fig. 02 — VA.2026.001.b a study of silence graphite· 29.7 × 42 cm · excellent exploring the shape of a pause — what happens between one thought and the next.
fig. 03 — VA.2026.001.c the first loop hand-drawn diagram· archival the earliest reasoning loop: think, hesitate, speak, sit with the result.
fig. 04 — VA.2026.001.d notes on uncertainty marginalia· variable · fragile marginalia from v0 development — every 'why' that became a design decision.
room 04 interface reviews sheet 04 / 09
fig. 05 — VA.2026.001.e a conversation in gray render silence preserved between turns — the deliberate gap where uncertainty lives.
fig. 06 — VA.2026.001.f the pause, rendered render a hesitation studied as an object — the system's most honest feature.
fig. 07 — VA.2026.001.g memory as a gallery render recalled experience, not storage — each memory hung in its own context.
room 05 catalog record sheet 05 / 09
materialsreasoning loop · long-term memory · dynamic mood engine · 24-hour timeline · multi-provider fallback · proactive messaging · group chat with auto-routing · sticker system · image analysisdimensionssource files — 100+ · database collections — 25+ · LLM providers — 8 · lib modules — 13+ · middleware layers — 5 · validation strategies — 3 · operational days — 37 · seed data — 6.10 MB · 4071 messages · 330 memories · 6251 shared memories
conditionterminated — v1 ended Aug 5 2026; v2 plannedacquired20 Mar 2026
revisionRev. Carchive locationDWG-001 · floor Floor 01
internal referenceVA.2026.001/Morigina personal impulse to give digital form to an imagined family
room 06 the archive sheet 06 / 09

failed experiments

✕ 2025

the perfect answer

taught it to be certain. it became a salesman.

✕ 2025

infinite fluency

removed every pause. it stopped feeling human.

✕ 2026

the hoard

let it remember everything. it drowned in itself.

✕ 2026

proactive hallucination

AI pretends to read user messages when none exist.

✕ 2026

timeline inconsistency

household members' schedules contradict each other.

✕ 2026

multi-LLM cost

2-7 calls per chat, estimated $0.02-0.08/chat.

✕ 2026

free provider unreliability

30s timeouts, rate limits.

✕ 2026

Cloudflare WAF blocking Dahl

MiniMax-M3 provider blocked by Cloudflare WAF on Vercel datacenter IPs.

draft concepts

  • a conversation with no deadline
  • the apology loop
  • an agent that refuses to answer
  • v0 prototype in Ionic Vue
  • Flutter mobile with 3 providers and 11 screens
  • LangGraph state machine for timeline generation

research notes

— uncertainty is not a failure mode — it is the medium itself.

— fluency without hesitation reads as performance, not thought.

— a human mind is a museum of hesitations, not a warehouse of answers.

— the seed data (6.10 MB) is not test data — it is a complete record of a personal AI companion system's interactions across its entire 37-day operational lifetime.

room 07 architecture · technical drawing DWG-001 · rev Rev. C
DWG-001 · Rev. C · floor plan — Void curator annotation — zones 03–05 are load-bearing; the presence layer must remain unshaded.
room 08 design statement sheet 08 / 09

I. Uncertainty as material

Certainty is easy for machines. The interesting material is doubt — the pause, the second thought, the sentence that ends without punctuation.

II. Slowness is a feature

A human-like mind is not measured in answers per second. The architecture values the silence between thoughts as much as the thought itself.

III. The goal is not a product

Void will not ship. It will be exhibited — unfinished, on purpose, for as long as the question deserves.

IV. Persona over utility

Characters reject questions outside their defined persona. The system deliberately avoids the helpful assistant default.

V. Time-aware reality

Characters know the current hour, day period, and gap since last interaction. They refuse to say good morning at 3 PM.

VI. Relationship consequence

Ignoring a character for 24+ hours decays affinity scores. A friend can regress to stranger over ~25 days of silence.

VII. Household ecosystem

Characters do not exist in isolation. They share a homeId, communicate via SharedMemory, and participate in group chats with auto-routing.

VIII. Cost-aware engineering

Every LLM call is logged with token counts. Cache-optimized prompt construction achieves 100% cache hit rate for the persona prefix.

IX. Graceful degradation

When LLM providers fail, the system falls back to personality-based rule generation. It produces a response even with zero LLM calls.

room 09 related works & future vision sheet 09 / 09

future vision

When Void is ready, it will not announce itself with a launch. It will simply begin remembering. Until then, this hall keeps the question open, and the lights low. v2 has not yet started — the database remains intact as a historical backup.