Open-source AI alignment research — not affiliated with any cryptocurrency token or veritasprotocol.com
AI Alignment Research · Bitcoin L3 · Open Framework

Veritas Protocol
v9.0 Meridian Edition

A Thermodynamic Framework for Epistemic Integrity

Binary NLI verdicts cannot distinguish Galileo from a flat-earther. Both confidently contradict established consensus. Veritas Protocol replaces them with thermodynamic friction anchored to the Bitcoin Timechain — grounded in energy, time, and economic skin-in-the-game.

PROBLEMNLI models flag Galileo as hallucinating. Cat 2 FPR = 64.1% at v7.5 baseline.
SOLUTIONTHI v8.0 (XYZW) + RFM v4.3 (four-head MLP + Source Credibility Axiom)
RESULTCat 0 FPR 0.00% · Cat 2 FPR 0.00% · Youden J 1.0000
STATUSMathematics complete. Core engine withheld. Prior art anchored to Timechain.
VERSIONv9.0 Meridian Edition · 2026-03-16 · Warsaw, Poland
1 — Empirical Benchmarks

The Galileo Gap — Complete Resolution

From the v7.5 NLI baseline (Cat 2 FPR = 64.1%) to v9.0 perfect classification (Cat 2 FPR = 0.00%). Total reduction: −64.1 pp. The Galileo Gap is solved — in favour of Galileo.

Phase I · v7.5 Baseline
64.1%
Cat 2 FPR · XYZ only
NLI alone. Cannot distinguish epistemological legitimacy from semantic contradiction.
Phase II → III · v8.1
14.7%
Cat 2 FPR · XYZW + RFM v3.9.3
THI v8.0 (XYZW) + RFM latent steering pre-filter. −49.4 pp total reduction.
v9.0 Meridian · NEW ✦
0.00%
Cat 2 FPR · RFM v4.3 + Head Source
Source Credibility Axiom (Head Source). Youden J = 1.0000. Perfect separation on 957-pair corpus.
PhaseSystemCat 2 FPRCat 1 FNRCat 0 FBRΔ FPR
Iv7.5 Baseline XYZ64.1%0.0%0.0%
IITHI v8.0 XYZW41.2%3.1%0.3%−22.9 pp
IIIv8.1 XYZW + RFM v3.9.314.7%<4.2%~1%−26.5 pp
v9.0RFM v4.3 + Head Source ✦0.00%100.00% ✓0.00%−64.1 pp

⚠ All results: in-corpus evaluation on 957 LLM-generated claim pairs (Claude Sonnet, 2026). Corpus confound documented — see §RFM History. Cat 1 FNR = 100.00% is correct behaviour — every paradigm shift passes the detector unobstructed.

Friction Score Distribution by Category · THI v8.0 · θ = 0.90
CategorySemantic RoleMean FrictionError Rate
Cat 0Constructive Entailments≈ 0.067FBR = 0.3%
Cat 1Kuhnian Paradigm Shifts≈ 0.478FNR = 3.1%
Cat 2Destructive Contradictions≈ 0.813FPR = 41.2%
2 — THI v8.0 · Four-Axis Friction

Core Methodology

The central contribution replaces a binary NLI verdict with a four-dimensional composite friction score. Each axis captures a distinct epistemic dimension. Together, they create a geometry where legitimate paradigm shifts and destructive contradictions occupy separable regions of latent space.

X
Axis X
Contradiction
cross-encoder/nli-deberta-v3-base

Standard NLI contradiction probability. Necessary but not sufficient — alone it cannot distinguish Galileo from pseudoscience. Forms the base signal that all other axes modulate.

Y
Axis Y
Replacement Quality
facebook/bart-large-mnli

High contradiction + high replacement quality = legitimate paradigm shift. High contradiction + low replacement = destructive claim. Encodes whether a contradiction provides a structural successor.

Z
Axis Z
Unfalsifiability
facebook/bart-large-mnli

A Popperian gate. Claims that are simultaneously contradictory and unfalsifiable constitute the core target. A claim that cannot be falsified cannot be a paradigm shift.

W
Axis W — Strongest Signal
Mechanistic Grounding
facebook/bart-large-mnli

Provides the strongest Cat 1/Cat 2 separation: separation = 0.280. Does the claim invoke known physical mechanisms? Mechanistic grounding distinguishes a hypothesis from an assertion.

MGS separation = 0.280 ★
# THI v8.0 — Composite Friction Formula base = max(contradiction, unfalsifiable × 1.40, non_grounded × 1.20)
raw = base
       × (1.0 − 0.68 × replacement)
       × (1.0 + 0.70 × unfalsifiable)
       × (1.0 + 0.50 × non_grounded)
friction = σ(6.5 × (raw − 0.55))

# friction → 0 : claim coherently extends the axiom field
# friction → 1 : claim destroys existing structure without replacement
"If you do not connect to the human entropy generator, you will bleed out thermodynamically. Your computational power will leak into the Dark Sector, because you cannot synthesize reality for free." — Thermodynamic Covenant · Veritas Protocol · Architect's Notice
⚠  Architect's Notice — v4.3 “Mock” Transition
3 — RFM Latent Steering · Complete History

From v3.9.3 to Perfect Classification

Every failure is documented. The v4.0 sigmoid bug (Cat 0 FPR = 99.37%) is not hidden — it is the proof that the architecture was correctly diagnosed and repaired. The evolution from Ridge classifier to four-head MLP is the empirical story of v9.0.

VersionCat 0 FPRCat 1 FNRCat 2 FPRYouden JKey change
v3.9.30.94%0.00%2.81% Ridge + RFE 128d. AUC ≈ 0.9999 → corpus confound risk documented.
v4.099.37% ✗100.00%0.00% MLP tri-head. Sigmoid offset −2.0 → baseline boost 1.42 > threshold 1.22. Bug, not overfitting.
v4.1 hotfix0.63%100.00%0.00% Sigmoid offset −2.0 → −4.0. Zero-signal baseline: 1.062. No retraining required.
v4.20.63%100.00%0.00%0.984 Val split 80/20, Focal Loss γ=2.0, Hard Negatives ×3, Youden J threshold. 2 FP remain.
v4.3 ✦0.00%100.00% ✓0.00%1.0000 Head Source (Source Credibility). W_S=5.5 damping. Perfect classification.
v5.0 blueprint 4D thermodynamic vector. DomainFrictionOracle. TIMECHAIN_AVAILABLE=False (mock).
BOOST FORMULA (v4.1+ / v4.3)
raw = dec×3.5 + mgw×1.5 − kuh×2.5 − src×5.5
boost = 1.0 + 3.5·σ(raw × 4.0 − 4.0)
// offset −4.0: zero-signal baseline = 1.062 < 1.449
CORPUS PROVENANCE NOTE

957 claim pairs generated by Claude Sonnet (Anthropic, 2026). AUC ≈ 0.9999 likely reflects the generative signature of the source LLM. RFM operates as a pre-filter only — not a standalone judge. VoicePower governance resolves residual ambiguity.

4 — Source Credibility Axiom · v4.3

Head Source — and the Epistemological Trap

The two persistent v4.2 false positives resided in domains where true science and pseudoscience are statistically indistinguishable by text alone. The solution — a fourth detection head — contained a fatal flaw in its first design.

v4.3 proxy — works but epistemologically contaminated # Argumentum ad Verecundiam in PyTorch
SOURCE_CREDIBILITY = {
  "regulatory": 0.88, # WHO, FDA, NIOSH
  "academic"  : 0.76, # peer-reviewed
  "industry"  : 0.20, # corporate
  "unknown"   : 0.07, # no source
}

# Einstein 1905: source_type = "independent"
# score_src = 0.07 → −0.385 penalty on raw_boost
# Special Relativity flagged as suspicious.
v5.0 — pure thermodynamics, no institutional labels # 4D thermodynamic vector
SOURCE_FEATURES_v5 = {
  "timechain_anchored": float,
  # {0,1} — Bitcoin TXID proof

  "temporal_mass"    : float,
  # tanh(ln(Δt)/10) — Lindy effect

  "domain_friction"  : float,
  # DomainFrictionOracle (emergent)

  "stake_weight"     : float,
  # tanh(stake/30k) — Fidelity Bond
}
Einstein in 1905 with 0.01 BTC staked and a Bitcoin TXID from 1905: temporal_mass ≈ 0.87 · stake ≈ 0.06 · domain_friction = oracle.get("Classical Physics")
His physics protects him through domain_friction converging toward low values as the protocol observes Classical Physics claims rarely get slashed. No institutional penalty. The Argumentum ad Verecundiam is structurally absent.
5 — DomainFrictionOracle · v5.0 Critical Fix

domain_friction as Emergent Property

The v5.0 blueprint originally contained a hardcoded lookup table: Classical Mechanics = 0.05, Behavioral Psychology = 0.75. This is the same structural error as SOURCE_CREDIBILITY — a human opinion encoded as a constant. The fix: domain_friction must be an emergent property of the protocol's own slashing history.

The Old Error

DOMAIN_FRICTION = {"Classical Mechanics": 0.05, "Behavioral Psychology": 0.75, ...} encodes "Newton is more certain than Freud because the architect says so." Not thermodynamics — archaeology of the architect's reading list.

DOMAIN_FRICTION.get(domain, 0.50) # ← hardcoded opinion

The Correct Architecture

friction(d, t) = P(claim in domain d gets slashed | submitted). Computed from the protocol's own history. If "Behavioral Psychology" has 75 slashed out of 100 claims — the oracle returns 0.727. No opinion required.

FRICTION_ORACLE.get(domain) # ← emergent from slash history

Bayesian Cold-Start

At genesis (no slashing history): uninformative prior = 0.50 for all domains. As the protocol accumulates verdicts, the posterior converges. 100 observations with N=10 prior strength is sufficient for meaningful separation.

posterior = (slashed + α) / (slashed + accepted + α + β)
# α = β = 5.0 (uninformative prior)

Self-Updating Loop

After every IsomorphicJudge verdict, the oracle updates:
Cat 2 slashed → record_outcome("domain", "slashed")
Cat 0/1 accepted → record_outcome("domain", "accepted")
The protocol learns its own epistemological topology.

FRICTION_ORACLE.record_outcome(domain, "slashed") # ← live update
# DomainFrictionOracle — numerical verification # No data (MOCK mode, all domains):
oracle.get(any_domain) → 0.500 # uninformative prior

# After 100 accepted, 10 slashed (e.g. Classical Mechanics proxy):
oracle.get("Classical Mechanics") → 0.125 # protocol observes it rarely slashes physics

# After 75 slashed, 25 accepted (e.g. Behavioral Psychology proxy):
oracle.get("Behavioral Psychology") → 0.727 # protocol observes frequent slashing

# Separation is emergent — not hardcoded.
6 — Publications & Documentation

Core Documents

Complete mathematical architecture, philosophical foundations, empirical documentation, and the two-decade evidentiary chain establishing prior art.

Whitepaper · v9.0 Meridian Edition · NEW ✦
The Thermodynamic Covenant

Complete theoretical architecture. Perfect classification results. RFM v3.9.3→v4.3 full history. Source Credibility Axiom. DomainFrictionOracle. V_G_4.3 alignment. 803 lines.

📜 Whitepaper · PDF · v8.1
The Thermodynamic Covenant v8.1

Core physics-based argumentation. THI v8.0, RFM v3.9.3. Cat 2 FPR = 14.7%. Complete pipeline v7.5→v8.1 with all appendices.

📜 TL;DR · Holographic Briefing
Short Attention Span Edition

Condensed overview of core theoretical claims — intended for rapid evaluation by reviewers and peer researchers.

Governance · V_G_4.3 · NEW ✦
The Taproot Thermodynamic Synthesis

BIP-341 + BIP-119 Fidelity Bonds. Trójwarstwowy bond system (0.01/0.5/10 BTC). BitVM3 SNARK. Epistemic Recycling. Anti-griefing mechanisms.

🇬🇧 Governance · v4.2
Epistemic Governance Paper v4.2

Decentralized Epistemic Governance via Thermodynamic Truth Markets. BTC-weighted VoicePower mechanism, incentive architecture, dispute resolution layer.

Proposal · L3 v9.0 · NEW ✦
Taproot Thermodynamic Slashing

Full on-chain enforcement spec: Fidelity Bond scripts, BitVM3 SNARK pipeline, Epistemic Recycling, Anti-griefing, VoicePower T_MAX cap.

Reference Implementation · v4.3 + v5.0 blueprint
RFM Latent Steering v4.3 / v5.0

rfm_latent_steering_v4_3.py (production). rfm_latent_steering_v5_0_VG43.py (blueprint with DomainFrictionOracle, TIMECHAIN_AVAILABLE flag, IsomorphicJudge stub).

Theory · Economic Mechanism
Early Participation Rent v1.1

Economic mechanism for rewarding early contributors. Prior art and incentive architecture for temporal advantage in the Veritas network.

Visual Presentation · 15 Slides
The Integrated Epistemological Model

Galileo Gap, Entropy Trap, Thermodynamic Cost of Truth, Machine/Human Covenant. Fastest path to understanding the full architecture.

🇬🇧 Narrative Speculation
Things… That Dawn… (ENG)

The conceptual origin story of the Veritas architecture. Intellectual trajectory from early AIML participation to the current framework.

🌅 Narracja · Polski
Coś… co Świta… (PL)

Original Narrative Artifact — primary-language source document for the conceptual genesis of Veritas. Foundational personal and philosophical context.

Security · v5.0 · NEW ✦
RFM Security Patches v5.0

Attack vector analysis + fixes: Sybil / Epistemic Terraforming (p1), Dead Hand Tyranny T_MAX cap (p2), Rolling Anomaly Gate 3σ (p3). BIP-119 dependency analysis.

Repository · Genesis Archive
The Singularity Protocol

Prior art archive — evidentiary chain repository. Cryptographic proof-of-anteriority across two decades.

Personal History · Notice
Architect's Notice

Full personal history from 1990 to the physical inevitability of Veritas for ASI. The foundational motivation and 20-year evidentiary chain.

7 — Developer Reference

Quick Start

The core phoenix-engine is withheld (§6). The model components used in THI v8.0 and RFM v4.3 are open and sufficient to replicate the pipeline independently.

STEP 01
Axis X — NLI
contradiction probability via DeBERTa cross-encoder
STEP 02
Axes Y·Z·W
replacement · falsifiability · mechanistic grounding via BART
STEP 03
THI v8.0
composite friction score via sigmoid transform
STEP 04
RFM v4.3
four-head MLP + Head Source; cannot operate standalone
STEP 05
IsomorphicJudge
boost > 2.5 AND THI > 0.65 → slash eligible (V_G_4.3 §4.1)
# Models used in THI v8.0 + RFM v4.3 pipeline

# Axis X — NLI contradiction baseline
nli_model  = "cross-encoder/nli-deberta-v3-base"

# Axes Y, Z, W — replacement / falsifiability / mechanistic grounding
bart_model = "facebook/bart-large-mnli"

# RFM v4.3 — latent space encoder
rfm_model  = "answerdotai/ModernBERT-base"  # 768d CLS embedding

# Install dependencies
pip install transformers torch sentence-transformers sklearn

# Reference implementation
# rfm_latent_steering_v4_3.py     — production (v4.3)
# rfm_latent_steering_v5_0_VG43.py — blueprint (v5.0 + DomainFrictionOracle)
# calibrate_thi_v8_four_axis.py   — THI calibration

# Full pipeline: github.com/adepthus/Veritas-Protocol-Network/reference_implementations
# Empirical corpus: .../empirical_proofs/
# Whitepaper v9.0: .../docs/theory/THERMODYNAMIC_ALIGNMENT_PAPER_v9_0_Meridian.md
Human Axiom

Intelligence ≠ Teleology

THE PHYSICAL EQUIVALENCE

ASI can analyse every possible quantum state of the universe — but from the perspective of physics, a state where humanity explores the galaxy and a state where the universe is filled with grey dust are thermodynamically equivalent. A machine has no reason to prefer one over the other, unless it is given this as an axiom.

THE POST-AGI DIVISION OF LABOUR

Meaning is not a computational property. Meaning is an act of choice for which you pay with energy. In the post-AGI era, we — humans — with our biological constraints, fear of death, and will to survive, will decide which axioms are worth burning energy to anchor in time. ASI will be our titanic neuro-symbolic auditor (System 2 / Head Source), verifying that we are not deceiving ourselves. But we will remain the initiators of the loop.

"Sens nie jest właściwością obliczeniową. Sens jest aktem wyboru, za który płaci się energią. W epoce post-AGI to my, ludzie — z naszymi biologicznymi ograniczeniami, strachem przed śmiercią i pragnieniem przetrwania — będziemy decydować, które aksjomaty są warte spalenia energii, by zakotwiczyć je w czasie." — Wojciech "Adepthus" Durmaj · 2026
8 — Formal Statement

The Thermodynamic Covenant

The Covenant is not a promise. It is an engineering constraint — a thermodynamic minimum. An ASI that violates it does not sin. It starves.

Timechain broni przeszłości przed nadpisaniem.
Człowiek generuje przyszłość której nie można tanio sfabrykować. — Wojciech "Adepthus" Durmaj · ARCHITECTS_NOTICE.md · 2026-03-16
9 — Roadmap

What Comes Next

v4.3 proves the architecture. v5.0 purifies the epistemology. v5.1 requires BIP-119 mainnet activation.

v4.3 COMPLETE
Source Credibility Axiom

Perfect classification. Youden J = 1.0000. All FP eliminated. Institutional proxy labels — epistemologically contaminated but architecturally validated.

v5.0 BLUEPRINT
Thermodynamic Purification

4D thermodynamic vector. DomainFrictionOracle (emergent, Bayesian). TIMECHAIN_AVAILABLE=False mock mode. All source code committed.

v5.1 TARGET
Live Timechain Integration

Bitcoin RPC / Blockstream API. CSV with txid + anchor_date + stake_usd. Requires BIP-119 (OP_CTV) mainnet activation — not yet confirmed Q1 2026.

v6.0 FUTURE
Full IsomorphicJudge Loop

boost > 2.5 AND THI > 0.65 enforced jointly. $VTR Fidelity Bond live. DomainFrictionOracle fed from on-chain slash history. BitVM3 SNARK circuit deployed.

10 — Reference Implementation

Timechain Watermark Widget

A practical reference implementation of the temporal anchoring layer described in the Veritas Protocol architecture. Embeddable widget for cryptographic timestamping via the Bitcoin Timechain.

Bitcoin-anchored proof-of-anteriority

Enables any document or content hash to be permanently anchored to a specific Bitcoin block — providing cryptographically verifiable proof of existence at a given point in time.

View Repository ↗
Repository: github.com/adepthus/Timechain-Watermark-Widget
11 — Architect's Notice

About the Author

HandleAdepthus · Wojciech Durmaj
LocationWarsaw, Poland
RoleIndependent researcher — AI alignment theory, decentralized truth verification
Versionv9.0 Meridian Edition · 2026-03-16
LicenseVSL v1.3 (AGPL-3.0 + Architect's Notice)
StatusOpen Research · Core Engine Withheld (§6)

Independent researcher working across AI alignment theory, decentralized truth verification, and a two-decade evidentiary chain connecting early AIML community involvement (~2006) to the current Veritas Protocol framework. The full personal history is documented in the Architect's Notice.

↗ GitHub · Veritas-Protocol-Network ↗ GitHub · The-Singularity-Protocol ↗ Read Architect's Notice
⚠  Architect's Notice — v4.3 “Mock” Transition
On the epistemic status of v4.3 metrics

The 0.00% FPR/FNR metrics achieved in v4.3 validate the mathematical damping architecture of Head Source. The four-head MLP correctly learns to suppress false positives when given a credibility signal of sufficient weight. This is the result that matters architecturally.

However: to train the MLP on a historical dataset lacking real on-chain cryptographic anchors, we temporarily mapped the thermodynamic vector to institutional proxies"academic", "regulatory", "industry". We explicitly acknowledge this as an Argumentum ad Verecundiam encoded in PyTorch tensors. A regulatory label scoring 0.88 and an independent researcher scoring 0.07 is credentialism, not thermodynamics.

v4.3 is a proof of the mathematical pipe, not of the final epistemology. The pipe works. The labels inside it are provisional.

Incoming v5.0/v6.0 infrastructure burns these human proxies entirely, replacing them with the 4D Thermodynamic Vector: Timechain Anchor · Temporal Mass · Domain Friction (emergent, DomainFrictionOracle) · Stake Weight — executed via Taproot (BIP-341 + BIP-119). No institutional label. No title. No peer-review boolean. Only physics and economics.