Michael Greshko
@michaelgreshko.bsky.social
Associate online news editor @Science. Freelance contributor to NYT, SciAm, WaPo, etc., and author of the Deviations newsletter. Former staff writer at National Geographic. Signal: mgreshko.01 https://linktr.ee/michaelgreshko
No: this article is a rehash of an article published by @livescience.com, for which I was interviewed. And no, I didn’t say anything to that effect with that level of conviction.
January 17, 2026 at 3:52 PM
No: this article is a rehash of an article published by @livescience.com, for which I was interviewed. And no, I didn’t say anything to that effect with that level of conviction.
Thank you! I’m not the first to ground the Voynichese generation problem in historical materials: Past studies have examined Cardan grilles and Alberti-inspired cipher disks/wheels, for example. But I wanted to put it all together with statistical modeling and a fully functional cipher system.
January 8, 2026 at 3:54 PM
Thank you! I’m not the first to ground the Voynichese generation problem in historical materials: Past studies have examined Cardan grilles and Alberti-inspired cipher disks/wheels, for example. But I wanted to put it all together with statistical modeling and a fully functional cipher system.
Hi there! Thanks for your interest! I’m the guy who came up with the cipher. Correct: I’m not claiming a full solution. Rather, the Naibbe cipher is meant to be a proof of concept, so future analyses of the manuscript can compare against a pseudo-Voynichese benchmark known to preserve meaning.
January 7, 2026 at 11:59 PM
Hi there! Thanks for your interest! I’m the guy who came up with the cipher. Correct: I’m not claiming a full solution. Rather, the Naibbe cipher is meant to be a proof of concept, so future analyses of the manuscript can compare against a pseudo-Voynichese benchmark known to preserve meaning.
Lo más importante es que estos tipos de estudios demuestran la posibilidad de una vía biológica para la autenticación de obras de arte.
January 7, 2026 at 4:58 PM
Lo más importante es que estos tipos de estudios demuestran la posibilidad de una vía biológica para la autenticación de obras de arte.
Rich Stone did a tremendous job writing and reporting this story: He has followed the Leonardo da Vinci DNA Project since the beginning.
You can check out Rich's reporting here: www.science.org/content/arti...
And the preprint itself here: www.biorxiv.org/content/10.6...
You can check out Rich's reporting here: www.science.org/content/arti...
And the preprint itself here: www.biorxiv.org/content/10.6...
Exclusive: Have scientists found Leonardo da Vinci’s DNA?
Inside the decadeslong quest to reveal the genes of a genius—and revolutionize art authentication
www.science.org
January 7, 2026 at 4:40 PM
Rich Stone did a tremendous job writing and reporting this story: He has followed the Leonardo da Vinci DNA Project since the beginning.
You can check out Rich's reporting here: www.science.org/content/arti...
And the preprint itself here: www.biorxiv.org/content/10.6...
You can check out Rich's reporting here: www.science.org/content/arti...
And the preprint itself here: www.biorxiv.org/content/10.6...
But if these researchers eventually pin down Leonardo’s Y chromosome, that portends a revolution in art authentication.
It'd be a coup for an emerging field called "arteomics," which analyzes trace biomolecules within an artwork to reconstruct its history and how it was made.
It'd be a coup for an emerging field called "arteomics," which analyzes trace biomolecules within an artwork to reconstruct its history and how it was made.
January 7, 2026 at 4:40 PM
But if these researchers eventually pin down Leonardo’s Y chromosome, that portends a revolution in art authentication.
It'd be a coup for an emerging field called "arteomics," which analyzes trace biomolecules within an artwork to reconstruct its history and how it was made.
It'd be a coup for an emerging field called "arteomics," which analyzes trace biomolecules within an artwork to reconstruct its history and how it was made.
It turns out that proving the DNA is Leonardo's is a really hard problem—because scientists can’t verify the sequences against any DNA samples known to have come from Leonardo himself.
His burial site was disturbed early in the 19th century, and he had no direct descendants.
His burial site was disturbed early in the 19th century, and he had no direct descendants.
January 7, 2026 at 4:40 PM
It turns out that proving the DNA is Leonardo's is a really hard problem—because scientists can’t verify the sequences against any DNA samples known to have come from Leonardo himself.
His burial site was disturbed early in the 19th century, and he had no direct descendants.
His burial site was disturbed early in the 19th century, and he had no direct descendants.
That's not proof that the "Holy Child" DNA is from Leonardo himself. Thousands of different people could potentially bear the same sequence in their Y chromosomes. The researchers say "it's a flip of a coin" if the DNA really belonged to Leonardo.
January 7, 2026 at 4:40 PM
That's not proof that the "Holy Child" DNA is from Leonardo himself. Thousands of different people could potentially bear the same sequence in their Y chromosomes. The researchers say "it's a flip of a coin" if the DNA really belonged to Leonardo.
In a new preprint, these researchers say they have recovered Y chromosome sequences from both this artwork and from a letter penned by a cousin of Leonardo.
Here's the wild part: Both sequences belong to a genetic grouping of people who share a common ancestor in Tuscany, where Leonardo was born.
Here's the wild part: Both sequences belong to a genetic grouping of people who share a common ancestor in Tuscany, where Leonardo was born.
January 7, 2026 at 4:40 PM
In a new preprint, these researchers say they have recovered Y chromosome sequences from both this artwork and from a letter penned by a cousin of Leonardo.
Here's the wild part: Both sequences belong to a genetic grouping of people who share a common ancestor in Tuscany, where Leonardo was born.
Here's the wild part: Both sequences belong to a genetic grouping of people who share a common ancestor in Tuscany, where Leonardo was born.
The drawing in question, "Holy Child," is disputed. Some say Leonardo did it. Others instead attribute it to one of the polymath's students.
Now the drawing is the focus of the Leonardo da Vinci DNA Project, a group of scientists on a decade-long odyssey to track down Leonardo's DNA.
Now the drawing is the focus of the Leonardo da Vinci DNA Project, a group of scientists on a decade-long odyssey to track down Leonardo's DNA.
January 7, 2026 at 4:40 PM
The drawing in question, "Holy Child," is disputed. Some say Leonardo did it. Others instead attribute it to one of the polymath's students.
Now the drawing is the focus of the Leonardo da Vinci DNA Project, a group of scientists on a decade-long odyssey to track down Leonardo's DNA.
Now the drawing is the focus of the Leonardo da Vinci DNA Project, a group of scientists on a decade-long odyssey to track down Leonardo's DNA.
Thanks! I've been in touch with René about it over the past year, and LiveScience got comment from him reacting to the study: www.livescience.com/archaeology/...
You can also see how the Voynich Ninja forum as a whole reacted to the original talk/preprint: www.voynich.ninja/thread-4848....
You can also see how the Voynich Ninja forum as a whole reacted to the original talk/preprint: www.voynich.ninja/thread-4848....
Mysterious Voynich manuscript may be a cipher, a new study suggests
A newly invented cipher may shed light on how the mysterious Voynich manuscript was made in medieval times.
www.livescience.com
January 6, 2026 at 8:49 PM
Thanks! I've been in touch with René about it over the past year, and LiveScience got comment from him reacting to the study: www.livescience.com/archaeology/...
You can also see how the Voynich Ninja forum as a whole reacted to the original talk/preprint: www.voynich.ninja/thread-4848....
You can also see how the Voynich Ninja forum as a whole reacted to the original talk/preprint: www.voynich.ninja/thread-4848....
In essence, the spaces between Voynichese "words" can't often be the actual spaces between unaltered plaintext words. Subject to that constraint, I'm totally cool with a homophonic substitution scheme that includes sigla abbreviations.
January 6, 2026 at 7:52 PM
In essence, the spaces between Voynichese "words" can't often be the actual spaces between unaltered plaintext words. Subject to that constraint, I'm totally cool with a homophonic substitution scheme that includes sigla abbreviations.
If so, then ciphertext tokens could often represent plaintext Latin words and interact with each other as such (e.g., as skewed pairs of words). I suspect a need for extra positional variability when encrypting continuous script: a faux prefix for -arum, a faux suffix for con-, etc.
January 6, 2026 at 7:33 PM
If so, then ciphertext tokens could often represent plaintext Latin words and interact with each other as such (e.g., as skewed pairs of words). I suspect a need for extra positional variability when encrypting continuous script: a faux prefix for -arum, a faux suffix for con-, etc.
I totally hear you. What I mean is, the devil is in the details of plaintext rebracketing. Even if you have multiple homophones for each sigla abbreviation, I worry that it would lend itself to a rebracketing scheme that's often neatly aligned with plaintext words' beginnings and ends.
January 6, 2026 at 7:33 PM
I totally hear you. What I mean is, the devil is in the details of plaintext rebracketing. Even if you have multiple homophones for each sigla abbreviation, I worry that it would lend itself to a rebracketing scheme that's often neatly aligned with plaintext words' beginnings and ends.
Thanks! The trick with sigla abbreviations within a putative VMS cipher, though, is that they couldn't be used all that often: Otherwise, Voynichese words would more consistently map cleanly onto Latin words and would thus behave much more Latin-like than they do.
January 6, 2026 at 5:58 PM
Thanks! The trick with sigla abbreviations within a putative VMS cipher, though, is that they couldn't be used all that often: Otherwise, Voynichese words would more consistently map cleanly onto Latin words and would thus behave much more Latin-like than they do.
Thanks, Lisa! To second Lisa’s bang-on summary, I’m not claiming a solution here—just trying to test whether and how it’s possible to reconcile something like Latin with the strange behavior of Voynichese.
January 6, 2026 at 5:07 AM
Thanks, Lisa! To second Lisa’s bang-on summary, I’m not claiming a solution here—just trying to test whether and how it’s possible to reconcile something like Latin with the strange behavior of Voynichese.
…You may find this study I recently wrote interesting: www.tandfonline.com/doi/full/10....
www.tandfonline.com
January 6, 2026 at 5:03 AM
…You may find this study I recently wrote interesting: www.tandfonline.com/doi/full/10....
Thank you! The short answer is, not without modifications—but it’s an interesting idea. For example, the previous card draw could determine the next respacing. In the 52-card variant (card values encode tables), red cards could be unigrams and black cards could be bigrams, yielding a 50-50 split.
January 5, 2026 at 2:40 AM
Thank you! The short answer is, not without modifications—but it’s an interesting idea. For example, the previous card draw could determine the next respacing. In the 52-card variant (card values encode tables), red cards could be unigrams and black cards could be bigrams, yielding a 50-50 split.
My point is that the information-theoretic properties of Voynichese, most notably its profoundly low conditional character entropy, place hard limits on how Voynichese can transmit meaning, if it contains meaning at all.
January 4, 2026 at 5:04 PM
My point is that the information-theoretic properties of Voynichese, most notably its profoundly low conditional character entropy, place hard limits on how Voynichese can transmit meaning, if it contains meaning at all.
To your point about whether it could be “channeled,” language-like text—in other words, glossolalia—that is another reasonable hypothesis for the text. The best recent version of this take comes from Joseph Fasano: semanticsarchive.net/Archive/2ZjN...
semanticsarchive.net
January 4, 2026 at 4:59 PM
To your point about whether it could be “channeled,” language-like text—in other words, glossolalia—that is another reasonable hypothesis for the text. The best recent version of this take comes from Joseph Fasano: semanticsarchive.net/Archive/2ZjN...
And there is also precedent at the turn of the 16th century for ciphers that give the appearance of being artificial languages, most notably Trithemius’s Polygraphia III, which was a major source of inspiration for the Naibbe cipher.
January 4, 2026 at 4:51 PM
And there is also precedent at the turn of the 16th century for ciphers that give the appearance of being artificial languages, most notably Trithemius’s Polygraphia III, which was a major source of inspiration for the Naibbe cipher.
Absolutely: People over the years have suggested the Voynich Manuscript may represent an early artificial language. It’s not historically out of the question; Hildegard de Bingen’s Lingua Ignota is ~200 yrs older. But the statistics of Voynichese are so odd, they constrain how a conlang can work.
January 4, 2026 at 4:48 PM
Absolutely: People over the years have suggested the Voynich Manuscript may represent an early artificial language. It’s not historically out of the question; Hildegard de Bingen’s Lingua Ignota is ~200 yrs older. But the statistics of Voynichese are so odd, they constrain how a conlang can work.