20 September 2026. No new verified plaintext. Three candidate words from the supplied Snow Crash passages were tested under 306 fixed settings. Inspection of the leading results found no coherent passage. Eighteen artificial recovery controls passed.
What the screenshots add
The supplied material covers 29 distinct displayed pages of Snow Crash and five of Cryptonomicon. Repeated screenshots of the same page are counted once. These are ebook display numbers, not stable print-edition page references. The attachment manifest records filenames, page assignments and file hashes. Full book pages are not reproduced here.
| Supplied passage | Observation in the novel | Relevance and limit |
|---|---|---|
| Snow Crash 266–268, 274–276 | Nam-shub connects language, incantation, code and changes to the mind. | Supports a thematic comparison. The fictional mechanism does not specify an LP cipher. |
| 270–272, 288, 293 | Cuneiform, a stylized tree identified as an asherah, and a law stele occur in the narrative. Later dialogue connects the tree and serpent with information. | A more specific visual comparison than generic ancient symbolism, but no distinctive geometry or ordered three-object match has been demonstrated in LP. |
| 291 | Exact written transmission is discussed as informational hygiene. | Resembles LP's emphasis on preserving words and numbers. Exact transmission also matters independently in cryptography. |
| 162 | Hacker vulnerability is explained through deep structures and raw bitmaps. | A fictional image/mind connection, not an instruction to invert an LP image or extract bits from whitespace. |
| 316–323 | Me and gis-hur are discussed through rules, commands and patterns, with computer-program analogies. | Motivates GISHUR as an exploratory key. It does not identify Sumerian as LP's plaintext language. |
| Cryptonomicon 584–587 | Pontifex uses an evolving permutation and modulo-26 addition; dialogue compares it with RC4 and discusses other permutation sizes. | Internal state size and output alphabet size are separate choices. A 29-symbol adaptation would require additional assumptions. |
| Cryptonomicon 588 | Surrounding wreck narrative continues. | No additional cipher algorithm was identified in this supplied page. |
These observations concern Stephenson's fiction. Characters' claims about ancient religion, language, medicine and historical identities have not been independently established by this reading. The passages do not establish Cicada's authorship, a Stephenson connection, or an early-Bitcoin connection.
Our earlier tree-image inspection did not identify a separate serpent, head or coil. Blur limits that observation: it is neither a positive identification nor proof that such a feature is absent. See Session 18.
Selected wording check
Re-ran the selected-term comparison against the retained community English translation. Eleven phrases and eight terms were queried, with “nam shub” appearing in both lists: 18 distinct queries, all with zero occurrences in 94 labeled translation entries.
Queries include informational hygiene, deep structures, neurolinguistic virus, self fulfilling fiction, code is just a form of speech, master of all the right commands, key words, patterns that rule the universe, perfect execution, gis hur, nam shub, Asherah, Enki, Enlil, Eridu, Babel, Hammurabi and Inanna.
Matching lowercases ASCII words, normalizes punctuation to spaces, requires whole-token sequences and does not cross entry boundaries. The 94 entries reflect the transcription format, not 94 LP pages. This is not a full-novel phrase comparison, and it says nothing about wording still encrypted in LP.
Fixed cipher experiment
The candidate list was fixed before this run's output inspection: NAMSHUB, ENKI, GISHUR. These are our exploratory selections, not keys uniquely specified by either book.
| Target | Runes | Selection prefix | Settings |
|---|---|---|---|
| LP2 page 32, tree page, archive image 49 | 121 | 72 | 102 |
| Section 7, starting within LP2 page 33 and continuing through page 39 | 1,589 | 952 | 204 |
| Total | 306 |
Rules:
- Zero-based 29-rune Gematria Primus indices. Each Latin letter in these three keys maps separately; SH is S then H. Key lengths are 7, 4 and 6.
- Test
c − k,c + kandk − c, modulo 29. - Test forward and reversed keys, with every cyclic starting phase.
- The tree page has one continuous counter. Section 7 has continuous and page-reset alternatives. Punctuation consumes no key positions.
- No added shifts, word resets, prime-valued keys, alternate spellings, F exceptions, autokey or extra layers.
The count is (7 + 4 + 6) × 2 directions × 3 forms × 3 target/counter combinations = 306. Settings need not be statistically independent.
Ranking uses the existing small LP1 English model on the first 60 percent, rounded down to complete words. The remaining text does not select candidates. It has been inspected during earlier work and is not a historically untouched holdout.
Controls and result
Two copies of known LP2 page 57 plaintext were artificially encrypted for each of three keys, three algebraic forms and two counter policies. All 18 controls recovered the entire plaintext at rank one. This checks recovery on constructed examples, not the unknown LP cipher or every possible language.
For the real tree page, the ten highest-ranked outputs were read in full. For section 7, the opening and ending 280 characters of each of the ten highest-ranked outputs were inspected. None yielded a coherent passage. Full outputs for all 306 settings remain in the results file; they were not all manually read.
The tree leader was NAMSHUB, forward phase 3, c − k. Section 7's leader was GISHUR, forward phase 2, page reset, c − k. These rankings are not positive findings: readable-looking fragments in otherwise incoherent output do not establish plaintext. The model may also misrank a real solution.
Decision and next step
Retain the Stephenson hypothesis as a literary comparison. This test supplies no support for these three repeated-key constructions on these two targets.
The Cryptonomicon passages make it reasonable to consider a modified permutation cipher, but they do not define one for LP. Before searching variants, require an independently motivated choice of output mapping, deck/state size, initialization and reset rule. Without that constraint, multiplying arbitrary adaptations would add little evidence. The standard Solitaire controls and the limitation of a simple 26-symbol-to-rune mapping remain recorded in Session 22.
The next useful discriminator is a distinctive, verifiable correspondence: an exact unusual phrase in solved LP, a matching illustration detail, or a structural rule that predicts an unexamined part of the ciphertext. More screenshots of similar themes alone would not strengthen a cipher claim.
Reproduce
- Probe code, run with Python and NumPy beside the existing core and transcription files.
- All results, source hashes and controls.
- Screenshot inventory.
The script performs no network requests. Its key choices and limits are explicit; neither novel's complete text is required.