MC-1 — the physical anchor artifact

Component deep-dive · 7 July 2026 · status: design. Ecosystem overview: /ecosystem.

Purpose

MC-1 is the physical unit of the anchor layer: a passive, write-once plate that carries a corpus release into environments that outlive institutions — orbits, the Moon, terrestrial vaults — and etches the release Merkle root so any recovered plate can authenticate any surviving digital copy. It is the disaster-recovery tier and terminal trust root of a living, catalog-indexed commons whose primary intended reader is a successor who inherits that catalog, not a stranger who stumbles on the object (decision 0002).

Current state

Status: design. No physical artifact exists; nothing has been mastered, qualified, or flown. What exists on disk is the concept spec (library/architecture/ecosystem.md, research rounds 1–2), the governing decisions (0004/0005/0006), and one unsent vendor inquiry (library/artifacts/mc1-nanofiche-inquiry-draft-2026-07.md), which PLAN.md marks awaiting approval as the next step and “the biggest open unknown.” The payload that would be etched already exists and verifies: Release 0000 (13 items, root cda2045f…3afd, glyph mc-r0000-0001). There is no quote, no master, no substrate decision, and no qualification plan.

Mechanisms / data model

Three-tier bootstrap, decode-taught-in-lower-tier. The design follows “Message in a Bottle” (Jiang et al. 2023): each layer is the key to the next, and the commons carries the full decode stack (decision 0002), so recovery never depends on a reader’s priors.

• Tier 0 — etched primer (naked-eye / low magnification). Self-describes with zero context: the neutral-hydrogen hyperfine time unit (1,420,405,751.768 Hz, Pioneer-plaque convention), epoch fixed by 13 pulsars to ~0.02-day accuracy (Weisberg & Paglen), a binary tutorial, and a catalog/map of the other artifacts. Medium: tungsten-in-Si₃N₄ or engraved nickel.
• Tier 1 — nickel Nanofiche analog library, readable with a magnifier (Rosetta precedent). Foundational set: math, physics constants, the genetic code, a Rosetta-style language sample, images of human life, music as notation + waveform, the chemistry of the senses (odorant molecules), and a charter of who we were. It also teaches Tier 2’s decode procedure pictorially.
• Tier 2 — dense digital wafer (fused silica / tungsten): the full curated corpus, unreadable without Tier 1’s instructions.

Materials, figures flagged. Nickel Tier 1: NanoRosetta verified ~737 pages/cm² annulus (~81,000 pages per 160 mm wafer) — about 3× below Arch Mission’s ~2,000/cm² claim; longevity “thousands of years” is a vendor figure with no test standard cited. Fused-silica 5D Tier 2: 360 TB/disc and “unlimited lifetime” are projections (2013 demo stored 3 bits/voxel, later ~4); it needs a polarization microscope. Tungsten-in-Si₃N₄: a ~1.8 eV Arrhenius barrier suffices for 1 Gyr and accelerated aging implies >1 Myr error-free — arXiv:1310.2961, preprint, extrapolated; its nested self-similar QR encoding is worth adopting. piqlFilm (film-copy option): 120 GB/reel, >1,000 yr vendor / ~750 yr @21 °C per Norner accelerated aging.

Form. ~100×100×3 mm, <300 g, fully passive — no power, no data line, no interface to the host bus, which is precisely what makes OEM bolt-on adoption plausible. The jacket is deliberately regular and engraved (gold-plated or ceramic) so a finder reads intent. The inquiry deliberately departs from Beresheet’s 40 µm foil stack (100 g, 25 nickel discs, NanoArchival — survival is an unfalsified self-assessment) toward a rigid ~1–3 mm standalone plate.

The etch — dual channel (decisions 0004, 0003). Every plate carries the release Merkle root twice: as the 64-character hex string engraved human-readably, and as the release-glyph SVG emblem (glyph for humans, code for machines — the two-channel rule, never one mark for both). The root is computed by corpus/tools/release.py: leaf = sha256 of raw file bytes; node = sha256 of the ASCII concatenation of two child hex digests; an odd node is promoted unchanged; items sorted by path. glyph_seed = int.from_bytes(bytes.fromhex(root[:8]), "big") = 3449947231 for Release 0000, seeding glyph mc-r0000-0001.

Placement (decision 0005), by the verified persistence ladder. GEO hosted payloads are the workhorse — century-scale peer-reviewed stability (~120–1,000 yr, Lewis et al. 2004); “billions of years” is Last Pictures publicity, not literature, so claims stay “millions credibly, billions plausibly”; place near the stable ~75°E/105°W libration points, avoid the 165°E/15°W equilibria. Lunar surface (~0.2 mm/Myr abrasion at cm scale) and lava-tube vaults (~290 K, <1 °C swing, >1,200 g/cm² basalt shielding — model outputs, no in-cave measurement, Horvath 2022). Escape trajectories for deep time; terrestrial mines/vaults for reachability. LEO is excluded categorically (FCC mission+5 yr cap). Redundancy over monuments: many cheap plates on whatever is already flying (KEO’s 30-year wait is the anti-pattern; Peregrine’s loss is the argument for copies).

Qualification & cost (thin sources, flagged). No qualification specs exist; the inquiry asks vendors only for temperature/humidity/radiation durability docs — full flight qualification (vibration, thermal-vac, outgassing) is an open path (inferred). Master cost is unpublished (the inquiry’s whole reason to exist). SpaceX rideshare is $350k base / 50 kg to SSO, $7,000/kg marginal (verified 2026-07-05, escalating ~$500/kg/yr), so a 300 g plate’s launch-mass cost is trivial — integration, qualification, and the host relationship dominate. “First GEO artifact ~$50–150k all-in” is UNVERIFIED round-1 analysis. Verified retail-funding precedent: Memory of Mankind €60/€150/€350; Celestis/LifeShip pricing remains unverified.

How it binds to the rest

• Commons (living layer): MC-1 carries a corpus release; the manifest and full catalog — including artifact locations — live online, so discovery is a lookup. Verification runs both directions: any artifact authenticates any digital copy; any two artifacts corroborate.
• Glyph system: release Merkle root → glyph_seed → the 3rdlayer-lights generator → the etched emblem. The generator spec must itself sit in the corpus (decision 0003) so the mark is reproducible from seed.
• AltersID (parked): corpus items carry DID-signed provenance; when keys and services are gone, the etched root is the terminal anchor.
• Aletheia layers: MC-1 is the anchor tier beneath Layer 3 (Library); crypto only has to survive between releases (renewal model, decision 0006), the etch survives after. Layer 4 (Gateway) is parked, and MC-1 has no online interface by design.

Key decisions

• 0002 — address successors, not strangers; MC-1 is the disaster-recovery tier and terminal trust root.
• 0003 — deterministic glyph paired with a machine channel; the etch is glyph + hex, never one mark for both.
• 0004 — releases are Merkle-rooted snapshots; the etched root is the terminal trust anchor.
• 0005 — placement GEO/lunar/escape/terrestrial; never LEO; redundancy over monuments.
• 0006 — integrity is active renewal (RFC 3161 now, SLH-DSA re-anchor before 2035); the etch is the last-resort anchor.

Open questions

• No vendor quote yet — master, per-unit, and rigid-substrate pricing all unknown; the inquiry is unsent.
• Full flight qualification undefined; only vendor durability docs are requested.
• GEO hosted-payload and lunar CLPS $/kg are unverified; only LEO rideshare is priced.
• The funding gap between a website and a first flown plate is open (a dedicated research round).
• Tier-2 encoding (bits/voxel, error correction, nested QR) is unselected; the mechanism by which Tier 1 pictorially teaches Tier 2 decode is unspecified.
• All media-longevity figures are vendor/projected — medium diversity is the mitigation, not proof.

What deeper builds next

1. Send the nanofiche inquiry (awaiting approval) to price MC-1 v0: a terrestrial prototype plate carrying Release 0000.
2. Produce the print-ready page package (Release 0000’s 13 items ≈ a few hundred rendered pages), the emblem SVG, and the 64-hex face layout.
3. Timestamp Release 0000 (RFC 3161 + OpenTimestamps) before any etch, so the etched root has a dated anchor (decision 0006).
4. From the quote, decide substrate (100 vs 160 mm; 1–3 mm rigid) and evaluate Arnano sapphire, MOM ceramic, and piqlFilm as alternates.