SOST: Proof of Irreversible Convergence. The complete technical specification covering ConvergenceX consensus, monetary policy, Gold Reserve, PoPC, governance model, security analysis, and constitutional rules. Version 4.5.
SOST treats «sound money» as four orthogonal properties that must hold simultaneously and forever. Two of them enforce credibility of issuance: the chain advances at a predictable rate (Time) under a measurable, energy-bounded effort (Work). The other two enforce credibility of value: total supply is finite (Scarcity) and an explicit fraction is converted, by consensus, into gold-backed reserves (Store of Value). All four are hard-coded at genesis with no governance override. Every section of this whitepaper either specifies one of these pillars or proves it.
Block production is paced at a 10-minute target via the cASERT unified rate controller.
Difficulty (bits_q) adjusts every block from a 288-block running average against
target, with bidirectional response: hardens when blocks come too fast, eases when too slow.
ConvergenceX equalizer profiles (E1–E7 relief, B0 baseline, H1–H20 hardening) are
mapped from chain lag and protect against runaway acceleration. The H20 ceiling and E7 floor
are immutable. From V11 onward a triangular cascade provides emergency relief on stuck blocks
without compromising the long-run schedule. See section 3 (cASERT) and section 12.
ConvergenceX is a native, original Proof-of-Work design. It is NOT a fork and NOT a parameter-tweaked variant of Bitcoin SHA-256, Monero RandomX, Ethereum Ethash, Equihash, CryptoNight, Scrypt, X11, BLAKE-anything, ProgPoW, RandomHash, or any other prior algorithm. A mining attempt requires solving a 32×32 SPD linear system through 100,000 sequential rounds of integer-only gradient descent and proving the solution sits in a stable basin of attraction (Proof of Irreversible Convergence). Memory-hard: 4 GB dataset (SplitMix64-indexed, state-dependent access from V11 Phase 1) plus 4 GB scratchpad per mining thread — 8 GB total. Per-block 256-op program changes every block. Verifiers reconstruct sampled rounds with ~500 MB of RAM in ~0.2 ms. See section 3.
The hard cap is 4,669,201 SOST — the first seven significant digits of
Feigenbaum's first constant δ (≈ 4.6692016091…), the universal
ratio that governs the period-doubling route to chaos in nonlinear dynamical systems.
The choice is deliberate: ConvergenceX itself is built on the same family of dynamical systems.
Emission follows an exponential decay per 131,553-block epoch with quotient
q = e-1/4 ≈ 0.7788; epoch 0 reward = 7.85100863 SOST.
No premine, no ICO, no dev tax, no minting function, no governance override.
See section 4 (Monetary Policy) and Appendix A (Canonical Constants).
Every block coinbase is split 50 / 25 / 25 by consensus
(PoW miner / Gold Funding Vault / PoPC Pool). Over the full emission curve this directs up
to 50% of total supply into two gold-backed channels:
25% — Perpetual Gold Vault. Protocol-funded, one-way TWAP
purchases of tokenised precious metals (XAUT / PAXG today, broader basket later) deposited
into a Heritage Reserve on Ethereum mainnet. Observable reserve ratio — not a peg,
not a redemption right. See section 5.
25% — Proof of Personal Custody (PoPC). Rewards distributed
to holders who prove continued custody of gold-backed positions through cryptographic
attestation. Models A (self-custody bonds) and B (escrow timelock) collateralise with
tokenised gold today and with physically self-custodied gold in the future, preserving
operator privacy on-chain. See section 6.
The split is enforced at validation: a block with the wrong coinbase shape is invalid.
See sections 5, 6, and Constitution C1–C15 (section 10).
Hardcoded at genesis · immutable forever · no governance override
A mathematically grounded Proof-of-Work based on verifiable dynamical-system certificates. The whitepaper covers the complete protocol specification including consensus algorithm, monetary policy, reserve architecture, custody protocol, and constitutional framework.
| Title | SOST: Proof of Irreversible Convergence |
| Version | v4.5 — March 2026 |
| Authors | NeoB (SOST Protocol) |
| Format | |
| Status | Production (mainnet) |
| Section 1 | Abstract |
| Section 2 | The Problem — PoW design space |
| Section 3 | ConvergenceX: Proof of Irreversible Convergence |
| Section 4 | Monetary Policy — Epoch-based emission with smooth exponential decay |
| Section 5 | Gold Protocol Reserve — Heritage vault |
| Section 6 | Proof of Personal Custody (PoPC) |
| Section 7 | Governance Model — No consensus governance |
| Section 8 | Economic Model — Demand, deflation, equilibrium |
| Section 9 | Security Analysis |
| Section 10 | The Constitution — C1–C15 |
| Section 11 | Roadmap |
| Section 12 | Technical Specifications |
| Appendix A | Canonical Constants (Normative) |
| Appendix B | Serialization (Normative) |
| Appendix C | PRNG Specification (Normative) |
| Appendix D | Test Vectors (Normative) |
| Appendix G | Emergency Reserve Authorization (V6 5-Defense Model) |
The whitepaper is published on multiple channels for redundancy and tamper-resistance. Verify the document hash against the published checksum.
| GitHub | github.com/sost-protocol/whitepaper |
| IPFS mirror | Pinned — immutable content-addressed copy |
| Website | sostprotocol.org/whitepaper |
| Checksum | SHA-256 published in repository |
| Name / Symbol | SOST |
| Genesis | 2026-03-15 18:00:00 UTC |
| Algorithm | ConvergenceX (4GB, 100k rounds, sequential) |
| Block time | 600 seconds target |
| Max supply | 4,669,201.609 SOST (hard cap) |
| Min unit | 1 stock = 0.00000001 SOST |
| Epoch | 131,553 blocks (~2.5 years, Feigenbaum α) |
| Initial reward | 7.85100863 SOST/block |
| Coinbase split | 50% miner / 25% gold / 25% PoPC |
| Annual decay | ~9.03% (smooth, no halvings) |
| Difficulty | cASERT bitsQ Q16.16, per-block; V5 (block 5,175+): avg288-based, dynamic cap (0%–3.0%); relief valve V8 single E7 cliff at 605 s (block 5,750+) → V9 staged drop 3/60 s from 540 s (block 6,550+) → V10 granular drop 1/60 s from 600 s (block 6,700+); historical: 24h half-life |
| Reserve | Heritage (sealed by default), Ethereum mainnet |
| Emergency | ≥95% miner signaling + 5-defense Gold Vault model (V6, block 10,000) |
| Premine / ICO | 0 — none |
| Consensus governance | None — immutable at genesis |
Lines-of-code totals for the SOST stack as of 2026-05-10, side by side with the Bitcoin Core reference implementation. LOC is a rough engineering metric — not a quality, security or decentralisation metric — but it gives a sense of how much surface area the project covers.
| Component | SOST | Bitcoin Core |
|---|---|---|
| Blockchain core (C++ / protocol) | 36,442 | ~280–310k |
| Front-end & explorer (HTML + JS + CSS) | 69,975 | — |
| Materials Engine (Python) | 161,282 | — |
| Geaspirit — geo-mineral ML (Python) | 9,386 | — |
| Auth gateway (Python) | 704 | — |
| Documentation (markdown) | 23,975 | (separate) |
| Total code (excl. docs) | ~278,000 | ~280–310k |
| Total code + docs | ~302,000 | — |
1. The blockchain core alone is 5–7× smaller than Bitcoin Core. Bitcoin Core has 15+ years of P2P hardening, scripting (Script + Taproot), thousands of tests, and dozens of network conditions handled in production. SOST does not, and pretending otherwise would be dishonest.
2. The full SOST stack matches Bitcoin Core in order of magnitude only because SOST is a vertical. The total includes a Materials Discovery Engine (161k LOC) and a geological prospectivity ML stack (Geaspirit) that Bitcoin Core does not contain. Comparing totals across different problem domains is not apples-to-apples.
3. What the numbers do show. SOST is a serious open-source project (~300k LOC total) with depth beyond a typical altcoin fork — the consensus is custom (ConvergenceX + cASERT), the wallet model is in-house, the front-end is hand-rolled, and the protocol is integrated with two external scientific systems.
4. What the numbers do NOT show. Network effect, peer count, security audits, deployed value, decentralisation. Those are independent of LOC and Bitcoin leads on every one of them.
Counts measured 2026-05-10 with find ... | xargs wc -l over each component's own source tree, excluding build artefacts and vendored code. Bitcoin Core figure is the upstream bitcoin/bitcoin master snapshot, code + tests, no docs.
The network and mining layers are already partially decentralized: multiple independent nodes, multiple active miners, open-source code, and no admin keys for block production.
However, two economic custody components are still temporarily centralized during Phase I (before V6 activates at block 10,000):
The protocol developer cannot change the constitutional 50/25/25 split, because it is hardcoded at genesis. What remains centralized during Phase I is custody, not the emission rule itself.
Custody transition roadmap:
The custody arrangement during Phase I is at the protocol developer’s operational discretion. The transition to protocol custody at block 10,000 has a hard date encoded in the consensus rules — it is not subject to the developer’s later willingness to step back. The Gold Vault must be able to change form, but not purpose.
SOST is not positioned as a Bitcoin clone. It combines Bitcoin-like timing discipline, CPU-heavy original work, mathematical scarcity, and a constitutional reserve design. The four tables below place SOST against six well-known PoW chains so the reader can see exactly where the design overlaps and where it diverges.
Snapshot date: 2026-05-08 UTC · Block heights and schedule drift change every minute and are intentionally rounded. Sources: Blockchain.info (BTC), BlockCypher (LTC, DOGE), localmonero.co (XMR), Blockchair (BCH, ZEC). SOST height read from the production explorer at the same time.
| Chain | Target / block | Height now | Expected | Drift | Status |
|---|---|---|---|---|---|
| Bitcoin | 600 s (10 min) | 948,511 | ~912,200 | +36,300 | Ahead ~250 days |
| Litecoin | 150 s (2.5 min) | 3,103,448 | ~3,063,000 | +40,400 | Ahead ~70 days |
| Dogecoin | 60 s (1 min) | 6,198,544 | ~6,531,800 | −333,300 | Behind ~232 days |
| Monero | 120 s (60 s before 2016) | 3,669,797 | ~3,672,000 | −2,200 | Essentially on time |
| Bitcoin Cash | 600 s | 950,193 | ~938,800 | +11,400 | Ahead ~80 days |
| Zcash | 75 s (150 s pre-Blossom) | 3,335,642 | ~3,349,000 | −13,000 | Essentially on time |
| SOST | 600 s | 7,713 | ~7,700 | +11…13 | Essentially on time |
Reading: chains with bulk retarget (BTC, LTC) accumulate large positive drift because their hashrate has grown exponentially faster than two-week recalibration can correct. BCH switched to ASERT in 2020 and the drift growth flattened. Dogecoin’s drift is negative because its merge-mining economy with Litecoin and DigiShield retarget combine to produce blocks slower than its 1-minute target. Monero, Zcash and SOST track their schedule within a fraction of a day — the property delivered by per-block retarget.
| Chain | Block reward (now) | Next change | Final supply | Notes |
|---|---|---|---|---|
| Bitcoin | 3.125 BTC | 2028 → 1.5625 | 21,000,000 BTC | Halvings every 210k blocks (~4 yr) |
| Litecoin | 6.25 LTC | ~2027 → 3.125 | 84,000,000 LTC | Halvings every 840k blocks (~4 yr) |
| Dogecoin | 10,000 DOGE (fixed) | never | No cap | Perpetual inflation, ~5.26 B DOGE / yr |
| Monero | ~0.6 XMR (tail) | fixed since 2022 | ~18.4 M + tail | Tail emission 0.6 XMR/block forever from June 2022 |
| Bitcoin Cash | 3.125 BCH | 2028 → 1.5625 | 21,000,000 BCH | Inherited BTC’s halving calendar at the 2017 fork |
| Zcash | 1.5625 ZEC | 2028 → 0.78125 | 21,000,000 ZEC | Halvings every 4 yr |
| SOST | 7.85100863 SOST (epoch 0) | block 131,553 (~2.5 yr) | 4,669,201 SOST | Continuous decay q = e−1/4 ≈ 0.7788 per epoch |
Bitcoin and its derivatives use discrete halvings. Monero and SOST use continuous decay. Doge does not decay at all (perpetual inflation). Monero’s tail emission means its supply has no asymptote — only a curve that flattens to a constant absolute issuance.
| Chain | PoW algorithm | Dominant hardware | Retarget | Notable design |
|---|---|---|---|---|
| Bitcoin | SHA-256d (double SHA-256) | ASIC | Bulk every 2,016 blocks (~14 d) | The original. Stateless flat hash. |
| Litecoin | Scrypt | ASIC | Bulk every 2,016 blocks (~3.5 d) | Scrypt holds a 128 KB working buffer; ASIC-friendly since 2014. |
| Dogecoin | Scrypt (merge-mineable with LTC) | ASIC via LTC | DigiShield (per-block) since Feb 2014 | Security tied to Litecoin’s merge-mined hashrate. |
| Monero | RandomX | CPU (deliberate) | LWMA (per-block, 720-block window) | CPU-oriented VM with JIT execution of randomised bytecode programs and a multi-MB-class dataset/cache plus per-execution scratchpads. ASIC-hostile by construction. |
| Bitcoin Cash | SHA-256d (same as BTC) | ASIC | ASERT (per-block) since Nov 2020 | Per-block since the post-fork DAA bug of 2017. |
| Zcash | Equihash ⟨200, 9⟩ | ASIC (since 2018) | DigiShield-like (per-block, EMA) | Originally GPU-friendly memory-hard; ASICs reached the chain by mid-2018. |
| SOST | ConvergenceX | CPU + 8 GB RAM | cASERT (per-block, avg288 + E/B/H equalizer profiles) | Native, original algorithm — not a fork. Solves a 32×32 SPD linear system across 100k sequential rounds of integer-only gradient descent and proves the solution sits in a stable basin (Proof of Irreversible Convergence). 256-op per-block program. 4 GB dataset + 4 GB scratchpad per mining thread. |
| Property | BTC | LTC | DOGE | XMR | BCH | ZEC | SOST |
|---|---|---|---|---|---|---|---|
| Hard cap | ✓ 21M | ✓ 84M | ✗ infinite | ✗ tail | ✓ 21M | ✓ 21M | ✓ 4.67M |
| Per-block retarget | ✗ bulk | ✗ bulk | ✓ DigiShield | ✓ LWMA | ✓ ASERT | ✓ DigiShield | ✓ cASERT |
| CPU-oriented PoW | ✗ | ✗ | ✗ | ✓ RandomX | ✗ | ✗ (was GPU) | ✓ ConvergenceX |
| Memory-hard footprint | ✗ | ~ 128 KB | ✗ | ✓ RandomX (multi-MB class) | ✗ | ~ partial | ✓ 8 GB total |
| Native (not a fork) | ✓ | ✗ | ✗ | ~ CryptoNight lineage | ✗ BTC fork | ✗ Equihash | ✓ ConvergenceX original |
| Constitutional reserve | ✗ | ✗ | ✗ | ✗ | ✗ | ✗ | ✓ 50% gold-linked |
| Default-private transactions | ✗ | ✗ | ✗ | ✓ ring + stealth | ✗ | ~ shielded opt-in | ✗ (E2E layer in DEX/Talk) |
| Hard cap derived from a math constant | ✗ | ✗ | n/a | n/a | ✗ | ✗ | ✓ Feigenbaum δ |
Comparison block last updated alongside the snapshot above. Tables refresh on each whitepaper revision.
Trinity is a vertical built on top of the SOST chain. The chain stays a strict Bitcoin-class PoW ledger; Trinity adds a discovery + proof + reputation layer above it. Two tracks run today: Earth Track (GeaSpirit-driven AOI proposals) and Materials Track (autonomous oxide candidate proposals). Both produce hash-bound proof bundles whose root SHA can be anchored on chain as a single capsule. The first anchor — Kalgoorlie Phase 1 at block #8085, txid d68678b5… — is the pipeline’s public birth certificate.
multi_ai_review.AICouncil reviews both Earth Track and Materials Track hypotheses; no per-track mock remains.SOST does not certify that a candidate material is revolutionary, that a candidate AOI hosts a deposit, or that any laboratory or field result has been obtained. Trinity bundles state the opposite explicitly: not experimentally validated, not DFT validated, not a patent claim, not a commercial performance claim. What SOST proves is something narrower but durable: this exact reproducible scientific process produced these exact documents at this exact block height, and they have not been altered since. That priority + traceability gives the work a public, immutable timestamp, which is the precondition for any future reputation, ranking or reward system the protocol may host.
One-line summary. Trinity is SOST’s first strong use case: turning autonomous digital research into publicly verifiable proof.
Full Trinity reference: sost-trinity.html · live proof registry on the public mirror.