Historical Knowledge Recovery

Historical Knowledge Recovery

Executive Summary

This document consolidates recovered technical knowledge from early Qubic development and related research efforts. Through systematic review of over 600 source files, 200 analysis scripts, and 100 data artifacts, we identified and categorized 15 major classes of previously undocumented or poorly documented findings.

The material spans several domains: architectural constants, cryptographic design constraints, economic subsystem specifications, early AI experiments, and statistical patterns in blockchain data. Each finding is assessed for evidentiary strength using the project's four-tier classification system.

Scope: This page merges and supersedes content previously spread across supplementary evidence notes and archaeological recovery logs. All findings have been re-evaluated for accuracy and assigned updated confidence ratings.

Methodology: Findings were extracted from script comments, output logs, Discord archives, forum posts, and code analysis. Each item was cross-referenced against at least two independent sources where possible.

Key Findings Summary

Section 1: Early AI Experiments and Theoretical Foundations

1.1 Ultima Online Agent System (2001-2007)

Between 2001 and 2007, developer Sergey Ivancheglo (known by the pseudonym "Come-from-Beyond" or CFB) conducted artificial intelligence experiments within the Ultima Online gaming environment. These experiments produced autonomous agents referred to as "Wisps."

Documented characteristics of the Wisps system:

From archived forum discussions (circa 2002-2003):

"The Wisps do not think in binary. They consider three states: advance, retreat, or observe. This allows for more natural behavior patterns."

Classification: Tier 3 -- the quote is attributed to CFB, but primary sources from this era are difficult to verify independently.

Relevance to Qubic architecture:

Limitation: A direct causal link between these early experiments and Qubic's architecture cannot be established from available evidence alone. The connection remains circumstantial.

1.2 The Aigarth Manifesto (2019)

The Aigarth paper sets out a theoretical framework for intelligence based on compression:

"Intelligence is the ability to find the shortest program predicting the system's state. The Priest seeks patterns; God creates randomness."

Core technical elements:

Connections to Qubic's design:

Temporal note: The paper was published exactly 11 years after the Pre-Genesis timestamp of September 10, 2008. The number 11 is Qubic's base architectural constant (11 squared = 121).

Classification: Tier 2 -- the document is publicly available and its contents are verifiable.

Section 2: Blockchain Forensics

2.1 Patoshi Pattern Analysis

The Patoshi pattern was first documented by Sergio Demian Lerner in 2013, identifying distinctive characteristics in early Bitcoin mining behavior.

Credits:

• Sergio Demian Lerner (@SDLerner) -- original pattern identification
• Jameson Lopp (@jlopp) -- data curation and public dataset

Dataset statistics:

Distinguishing characteristics:

Distribution across Qubic computors (not confirmed):

Earlier versions reported that 2,835 Patoshi block heights spread over 676 bins with a standard deviation of 0.05 and a coefficient of variation of 1.2%, and read this as structured allocation. These values are impossible: for any split of 2,835 whole blocks into 676 bins, the standard deviation is at least 0.395 and the coefficient of variation at least 9.4%. The reported figures are retired; the reading was not confirmed.

2.2 Hex-Word Block Spacing

Analysis of hexadecimal-readable words occurring in Patoshi block hashes revealed a pattern of exact 27-block spacing between semantically meaningful hex strings.

Hex-word vocabulary searched:

HEX_WORDS = [
    'beef', 'dead', 'cafe', 'face', 'fade', 'feed',
    'deaf', 'dada', 'bead', 'abba', 'acdc', 'abbe',
    'babe', 'deca', 'bade', 'abed', 'cede', 'deed',
    'decade'
]

Selected 27-distance pairs:

A total of 20 pairs were identified with exactly 27-block separation and semantically meaningful hex words on both ends.

Statistical context:

Hex-word occurrences: 771 in 21,953 Patoshi blocks
Total 27-distance meaningful pairs observed: 20

Classification: Tier 3, hypothesis. The word list, the spacing of 27 and the notion of "meaningful" were chosen after looking at the data, and no null model was tested. With 771 hex-word occurrences, many pairs at any fixed distance are expected by chance.

2.3 The "DEAD" Block Pattern

Four Bitcoin blocks were identified containing the hex string "DEAD" in their public key data, each divisible by 27, and all remaining unspent:

Under a naive uniform-random null (each property assumed independent), the joint probability of four blocks sharing these three properties (hex "DEAD" presence, divisibility by 27, unspent status) is estimated at roughly p < 10⁻⁸. This figure is not Bonferroni-corrected for the space of 4-hex pattern words (there are 16⁴ ≈ 65 k possible 4-hex combinations that could have been searched) or for the divisor pool, so the effective multiple-testing burden is substantial. The pattern is real and on-chain, but the bare p-value should be interpreted as suggestive rather than conclusive.

Classification: Tier 2 -- data is verifiable on-chain, though causal interpretation remains open.

Section 3: Cryptographic Design Analysis

3.1 Position 27 Constraint in Qubic Identities

Position 27 in Qubic identity strings is constrained to only four values (A through D), while symmetric positions accept eight values (A through H):

byte_15_max = 127
mathematical_proof = "floor((2^63 - 1) / 26^13) = 3 = D"

The constraint follows from how Qubic encodes identities: the 32 public-key bytes are split into four little-endian 64-bit values, each written as 14 base-26 characters. The last character of each group can only take a few values, because floor((2^64 - 1) / 26^13) = 7; for a fragment limited to 63 bits the bound is 3. It is a property of the encoding, not a design pattern.

Classification: Tier 1. The derivation is verified and the constraint is deterministic.

3.2 The 28/27 XOR Anomaly

Identity analysis revealed a persistent off-by-one discrepancy in XOR operations:

Expected XOR result: 27
Actual XOR result:   28
Difference:          1 (Hamming distance: 3 bits)

Transformations that reconcile the discrepancy:
  28 XOR 7 = 27
  28 - 1   = 27

Bidirectional properties:

28 XOR 56 = 36
28 XOR 36 = 56

Both 36 and 56 appear independently as significant values in other analyses.

Possible explanations:

• The encoding scheme includes a subtle modifier (XOR with 7, or subtraction of 1)
• Character encoding at position 27 differs by one unit
• A multi-stage processing pipeline applies an additional operation

Classification: Tier 3 -- the anomaly is reproducible but its cause is not yet determined.

3.3 The prevDigest Chained Hashing Method

CFB provided a specific cryptographic hint in community discussions:

"prevDigest is supposed to reveal Salted value"

Implementation approach documented in analysis scripts:

def prevDigest_method(data):
    """
    K12 hash of previous digest concatenated with current data.
    The hint suggests chained hashing reveals a salt value.
    """
    return k12_hash(prev_digest + data)

Status: This method was implemented and tested but has not produced the expected result. The approach is documented here for completeness and future investigation.

Classification: Tier 3 -- the hint is attributed to CFB, but the method remains unvalidated.

Section 4: Protocol Economics and Architecture

4.1 The Qus/Paracosm Economic System

Discord archive analysis recovered specifications for an internal economic subsystem that was not included in public documentation at the time of writing:

Key statements from Discord archives:

• "Qus need to be burned for smart contracts to run"
• "Qus will be used to run everything related to Paracosm on Qubic"
• "Sacra are used within paracosms, Qus outside"

This two-tier economic model (Qus for the base layer, Sacra for application layers) represents an early design for resource allocation in a decentralized computing network.

Classification: Tier 2 -- statements are preserved in archived Discord logs with timestamps.

4.2 The ZgZ Address Steganography Hypothesis

The Bitcoin address 1CFBdvaiZgZPTZERqnezAtDQJuGHKoHSzg contains the substring "ZgZ". On 20 March 2023 CFB wrote on Discord (message 1087514727179104307), with a trollface emoji:

"g in ZgZ will prove that 1CFB2 is not random, and that will be pretty convincing proof, so it's not that easy"

The trollface suggests it may have been meant as a joke. Readings that have been proposed include:

  ZgZ = Zigzag reading pattern
  - Read matrix data diagonally
  - Alternate direction per row

Additional linguistic component:

• The substring dva is the Russian word for "two"
• dvaZgZ could be read as "two zigzags"

Classification: Tier 3, hypothesis. The readings are chosen after the fact and have not produced any testable result.

Section 5: Recurring Constants and Number Theory

5.1 The CFB Constant Set

A recurring set of numerical constants appears across multiple independent analyses:

CFB_CONSTANTS = [7, 27, 47, 121, 137, 283, 676, 817]

5.2 Extended Constants Collection

An extended set compiled from multiple analysis scripts:

EXTENDED_CONSTANTS = [
    27,    # 3^3, ternary cube
    121,   # 11^2, architectural constant
    576,   # 24^2, grid mathematics
    676,   # 26^2, computor count
    969,   # Palindrome
    2187,  # 3^7, transaction size parameter
    2673,  # Transaction size parameter
]

Notable observation -- Block 576:

Block 576: ExtraNonce byte = 0x1b = 27 (decimal)

This provides a concrete on-chain instance where 27 appears embedded within a block whose height (576) is itself a perfect square (24 squared).

Classification: Tier 2 -- the constants are objectively present; their intentionality is inferred.

5.3 Fibonacci References

Multiple analysis scripts reference Fibonacci sequences as potential encoding keys:

fibonacci = [1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144]

The number 21 appears with particular frequency across analyses. It is simultaneously a Fibonacci number, the chapter number in the Revelation 21:6 reference found in several scripts, and the product of 3 and 7 (both members of the CFB constant set).

Classification: Tier 3 -- the references are documented but their functional role is uncertain.

5.4 Helix Pattern Excess in the Anna Matrix

The Anna Matrix contains 4.9 times more "Helix patterns" (triplets whose values sum to a multiple of 3) than expected from a random distribution:

This result is reported from an older analysis and was not re-run in the September 2026 re-check; the random baseline used is unclear. It does not support a design conclusion. The matrix is a hand-compiled table of bot replies.

Section 6: Code Style and Authorship Indicators

6.1 Stylometric Observations

Analysis of early Bitcoin source code identified several stylistic markers:

Quadruple-slash comments:

//// issue here: it doesn't know the version.
//// is this all we want to do...
//// or maybe

Plural pronoun usage:

// We need to...
// We should...

Integer-first arithmetic philosophy: A consistent preference for integer arithmetic over floating-point operations, matching CFB's documented coding practices.

6.2 Comparison with CFB's Known Projects

Classification: Tier 3 -- stylometric analysis is inherently uncertain. Many experienced programmers share similar conventions, and coding style can be consciously altered. This evidence is suggestive but not conclusive.

Section 7: Negative Results and Failed Approaches

7.1 Systematic Exclusion Catalog

A total of 1,187 distinct extraction and derivation methods were tested and produced no valid results:

Interpretation: The systematic failure of all tested methods suggests one or more of the following:

• A missing transformation step that has not yet been identified
• An unknown salt or key component not present in public data
• A time-dependent mechanism that prevents premature extraction
• A fundamentally incorrect analytical framework

These negative results are documented here because they constrain the space of viable hypotheses and prevent duplication of effort by future researchers.

7.2 Untested High-Confidence Candidates

Nine seed candidates were identified with estimated 85-95% confidence ratings but remain unvalidated:

These candidates are recorded for future validation efforts.

Section 8: Esoteric and Cultural References

8.1 Revelation 21:6 as a Cipher Parameter

Multiple scripts reference the biblical verse "Revelation 21:6" as a cryptographic parameter:

"I am Alpha and Omega, the beginning and the end."
  -- Revelation 21:6

Numerical interpretation:
  ALPHA = 6 (verse number)
  OMEGA = 21 (chapter number)
  Sum: 21 + 6 = 27

Additional properties:
  Reversed: 6:21
  Ratio: 21 / 6 = 3.5 (half of 7)
  Product: 21 * 6 = 126 = 2 * 63 = 2 * 7 * 9

Classification: Tier 3 -- the reference is present in multiple scripts, but its functional role in any cryptographic scheme is unconfirmed.

8.2 The LOST Numbers Sequence

The sequence [4, 8, 15, 16, 23, 42] (from the television series "LOST") appears in forensic analysis code:

LOST_NUMBERS = [4, 8, 15, 16, 23, 42]  # Sum = 108 = 27 * 4

The sum of 108 equaling four times 27 connects to the broader constant framework, though this may simply reflect a cultural reference embedded by the developer.

Classification: Tier 3 -- the presence is documented; significance is uncertain.

Source Assessment

Evidence Quality by Domain

Source Materials

Limitations

Methodological Constraints

1. Source verification gaps: Historical material from the 2001-2007 era relies on forum archives that may be incomplete or inaccurately attributed. Independent verification of early Ultima Online experiments is not currently possible.

2. Interpretive subjectivity: Several findings (hex-word semantics, address steganography, cultural references) require interpretive judgment. Different analysts may reach different conclusions from the same data.

3. Survivorship bias: The recovered artifacts represent what was preserved in scripts and logs. An unknown quantity of additional context may have been lost, discarded, or never recorded.

4. Causal ambiguity: Statistical correlations, even when strong, do not establish causation. The recurring appearance of specific constants could reflect intentional design, mathematical necessity, or selection bias in analysis.

5. Incomplete negative results: While 1,187 methods were systematically tested, the space of possible approaches is effectively unbounded. The negative results constrain but do not eliminate viable hypotheses.

Alternative Explanations

For each major finding category, at least one plausible alternative explanation exists:

Recommendations for Future Work

1. Independent replication: All reproducible findings should be verified by researchers outside the original analysis team.
2. Primary source recovery: Efforts to locate and archive original Ultima Online forum posts and early IOTA/Qubic development logs would strengthen the historical record.
3. Seed candidate validation: The nine untested high-confidence candidates documented in Section 7.2 represent the most actionable open items.
4. Expanded stylometric analysis: Application of formal authorship attribution methods (rather than informal comparison) would improve the rigor of code style findings.
5. Temporal analysis: A systematic study of date-aligned events across all known CFB publications could either confirm or refute the pattern of anniversary timing.

References

• Lerner, S. D. (2013). "The Well Deserved Fortune of Satoshi Nakamoto, Bitcoin creator, Visionary and Genius." Blog post.
• Ivancheglo, S. (2019). "Aigarth." Published September 10, 2019.
• Lopp, J. Public Patoshi block dataset. Accessed January 2026.
• CFB Discord archives. Qubic community server, various dates.
• Analysis scripts: scripts/forensic_analysis/, scripts/decode_27_message.py, analysis/investigate_position27_anomaly.py, and related files in the research repository.