David F GliddenandClaude Opus 4.8
3f9a89b00c
Basic Memory v0.21.6 first sync over the live memory dir (steward-authorized live-dir trial, Option A 2026-06-06): adds permalink: to frontmatter, refolds long YAML description lines, strips final newlines. Bodies untouched — verified via full diff classification. From this commit forward, any diff in claude/memory shows only what Basic Memory or the session writes. Trial design: MemPalace untouched as incumbent; git status check on this dir at every wrap; end-of-day evaluation (recall quality, sync robustness, rebuild-from-files, malformed-file behavior). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
30 lines
3.1 KiB
Markdown
30 lines
3.1 KiB
Markdown
---
|
|
name: MemPalace's role — instrumentation for steward+Claude collaboration during BMF
|
|
L1-degraded period
|
|
description: MemPalace is not a deliverable, not a long-term substrate, not part of
|
|
CapableMind. It is a working tool the steward chose to instrument our collaboration
|
|
while BMF L1 reliability is being fixed. Holds 100% verbatim recall locally so that
|
|
working memory across sessions is not dependent on BMF.
|
|
type: project
|
|
originSessionId: 597aba4e-942b-4f20-8a89-3d801b14c971
|
|
permalink: claude-memory/project-mempalace-role-as-instrumentation
|
|
---
|
|
|
|
# MemPalace's role in the current period
|
|
|
|
Steward verbatim 2026-05-01: *"Seb has nothing to do with mempalace — I am using it to help us while we finish the BMF L1 problems. Mempalace is its own project and I am using it to help me to help you to help me."*
|
|
|
|
**Why:** BMF L1 has been unreliable through April (deep-audit work since 2026-04-19; latest pathology surfaced 2026-04-30 — 1,349 index segments + 25k transactions in lance store, NOT issue #10). While L1 is being repaired, the steward needs a memory substrate that *works now* — verbatim recall, local, no API, no cloud dependency. MemPalace is that stand-in. The chain "I am using it to help me to help you to help me" is precise: steward → MemPalace → executor's working context → steward's continuity.
|
|
|
|
**How to apply:**
|
|
- **Don't conflate territory.** MemPalace is its own open-source project (Yousef Amar et al.). It has its own contributors, PRs, issues, design philosophy (`~/_Dev/mempalace/CLAUDE.md`: "Memory is identity… 100% recall is the design requirement"). It is not BMF, not L2, not CapableMind. Seb is not in the loop.
|
|
- **Don't propose MemPalace as a permanent substrate.** It is the right tool *now* because L1 isn't ready. When BMF L1 reaches reliability, the architectural question of how MemPalace and L1 relate is a separate later decision — not for the executor to pre-empt.
|
|
- **Treat MemPalace work as enabling work, not deliverable work.** Time invested in MemPalace transition (bge-m3 + MPS, sub-room granularity, future updates) is overhead in service of the actual work — ARC, L2 governance docs, BMF L1 audit support. The steward's 2026-05-01 *"I almost regret doing"* about the bge-m3 transition is a signal that the overhead has felt disproportionate; keep transition work proportionate to its instrumental purpose.
|
|
- **Upstream MemPalace conversations** (PRs, issues, contributors) belong to MemPalace's own ecosystem — never refer as "Seb's network" or similar BMF-orbit phrasing. See `feedback-arc-territory-steward-and-claude-code-only.md` for the broader territory rule.
|
|
|
|
**Operationally for current period (through bge-m3+MPS transition completion):**
|
|
- Chamber-library wing first (in flight 2026-05-01)
|
|
- Phases 4b/c/d (Obsidian vault, Claude transcripts, CapableMind thinking) sequential overnight runs after
|
|
- Discipline changes deferred until mine completes per steward 2026-05-01
|
|
|
|
When the transition completes and MemPalace is stable on the new substrate, the question shifts back to BMF L1 audit work + the live ARC/L2 work — MemPalace returns to background instrumentation. |