Zakhor

Letter no. 6 · Monday, September 7, 2026

Hello,

Zakhor, the week — what two million names make possible

Last week, we asked why memory is a commandment. Why zakhor — imperative, second person, no escape — and not "remember if you can," "remember if you have time." The answer lay in one idea: because a people that ceases to remember ceases to exist before it even disappears.

This week, the question becomes more troubling, and more beautiful: what if we could truly obey?

Five months. Two million people identified, sourced, connected. This number has no precedent, and it is not an ending — it is the moment when we discover that the tool holds, that the method scales, and so we must seriously ask ourselves the question we did not dare to: how far?

So let us go. Here, without excessive caution, is what Zakhor can become.

I. The Census: 120 Million

Since Abraham, approximately one hundred twenty million people have lived and died as Jews. This is a figure of magnitude, not a precise count. But it is the only horizon worth considering: not "a large database," but the complete memorial census of a civilization.

No generation before ours could have formulated this sentence without appearing ridiculous. We can, for a precise reason: the three deposits bearing these names are now readable by machine.

Administrative documents. Marriage certificates, ketubot, civil registers, taxpayer lists, community rolls, consular registers. The bureaucratic fabric of centuries — that which, for once, serves us.

Documents of memory. Narratives, correspondence, memorials, yizkor bikher, rabbinic chronicles, wills. What families wrote about themselves.

Stones. This is our most neglected deposit, and doubtless the richest. Millions of gravestones bear, in Hebrew, in abbreviations, in acrostics, in chronograms, the complete identity of the deceased: his name, that of his father, his date, often his priestly lineage, sometimes his trade and character. These inscriptions have been abandoned because deciphering them required a paleographer per cemetery. This is no longer true. Recognition of handwritten and lapidary Hebrew script has crossed, in the past two years, the threshold separating curiosity from working tool. Every Jewish cemetery in the world is now a database awaiting reading.

Crossing these three sources is the heart of the method. A name attested by a document, confirmed by a stone, illuminated by a narrative: here is an established fact. A name borne by a single source: here is a conjecture, and it will be stated as such.

II. The standard: established, conjectured, transmitted

This is our most important decision, and the least spectacular.

Every piece of information entered into Zakhor carries its epistemic status: historical, memorial, established, conjectured, transmitted. A family tradition tracing the lineage back to Rashi is not false — it is transmitted, and it is precious data in itself, about what a family believed and told of itself. It is simply not of the same nature as a notarial deed from 1704.

This requirement is not academic affectation. It is a necessity for the project's survival. The major risk of any AI-assisted memory is the feedback loop: a conjecture enters the corpus, a synthesis takes it up, a second synthesis cites the first, and three iterations later a hypothesis has become a fact, without anyone having lied. The day this happens at scale, Zakhor is worthless.

The most recent models finally give us the means to hold this line: they know how to qualify the reliability of what they assert, distinguish what they know from what they infer, and refuse to fill a void with plausibility. We are building as the university builds: on verified works, each refutable, each traceable to its source. This is how a closed circle becomes a shareable truth.

III. The Great Projects

The Conversos

Five hundred thousand people forcibly converted over five centuries in the Iberian peninsula and its empires. Approximately ten million living descendants today, the vast majority unaware of this ancestry — in Portugal, Spain, Brazil, Mexico, Peru, in the southwest of the United States.

This project is within our reach for a paradoxical reason: the Inquisition was meticulous. It produced, in order to persecute, the most detailed genealogical corpus of the premodern world — trials, genealogies of the accused, limpieza de sangre, nominative denunciations spanning three generations. Researchers from the University of São Paulo have been working for twenty-five years on these digitized archives. Contact is currently being established.

To turn the machinery of the Inquisition against its own purpose, and to restore to ten million people a name that had been torn from them: this is perhaps the most beautiful historical project of the decade. We want to undertake it.

The Manuscripts

Hundreds of thousands of folios in Hebrew, Judeo-Arabic, Judeo-Spanish, Judeo-Persian, Aramaic, lie dormant in the collections of Saint Petersburg, Cambridge, Jerusalem, New York, Cairo. The Geniza alone contains more documents than a researcher's lifetime can read.

Transcription, alignment, translation, open publication. Not to replace Hebraicists, but to return to them the time that decipherment was stealing from them, and to open to everyone texts that fifteen people in the world knew how to read. The site is designed in thirteen languages, with management of transliterations and onomastic variants — for the same man is called Mordekhaï at the synagogue, Marcos at the registry and Mordoche on his tomb, and knowing that it is the same man is already all the work.

Genealogy for All

Tracing two or three centuries of one's own lineage costs thousands of euros today for those who go through a professional. We want it to be free, for every Jewish family in the world, and to be accomplished in a matter of hours.

Not because it is a nice service. Because it is the condition for the corpus to fill: each family that receives its Grand Livre corrects it, enriches it, contributes a photograph, a letter, a missing name. Generosity is not here a moral posture — it is the very architecture of the project.

Genetics

This is the new project, and the one that demands the most rigor.

The Y chromosome is a chronometer. Among the approximately seven hundred thousand markers tested in high-resolution analysis, we observe about one mutation every eighty years. This regularity makes it possible to date the most recent common ancestor between two men with a precision on the order of a generation, and thus to validate — or refute — a documentary hypothesis. A tradition linking two Cohen branches separated since Fez and Livorno becomes testable.

Mitochondrial DNA, transmitted by mothers, is a different instrument: its mutations are spaced roughly four hundred years apart. It permits neither fine dating nor precise geographic localization. Saying this clearly is part of the work: we will not promise what the marker cannot give.

And there is the medical dimension, which is not anecdotal. Jewish lineages carry identifiable pathological signatures — collagen pathologies are notably frequent in North Africa; a mutation of the COL4A4 gene, associated with Alport syndrome, is traced through specific families. Whether a transmission is autosomal or sex-linked changes everything for the descendants concerned. A well-constructed genealogy is an act of public health.

Two absolute rules, and they will not budge: nothing genetic enters Zakhor without the explicit, informed and revocable consent of the person; nothing leaves it toward an insurer, an employer or an administration. A geneticist who does not set this limit before opening the laboratory has understood nothing of what we are.

The Living Encyclopedia

The Encyclopaedia Judaica was the monument of the twentieth century: thirty-two volumes, frozen the day they were printed. Zakhor is its dynamic successor — a corpus interrogable by conversational agent, which adapts to who interrogates it. The historian, the sociologist, the rabbi, the adolescent seeking the origin of his name do not enter by the same door and do not receive the same answer. The collection itself is the same, and it grows each night.

This opens questions that no one could ask until now. How did marriage customs evolve among religious Jews over three centuries? What is the real geography of Sephardic merchant networks? How do surnames become deformed at the passage of a frontier, a language, a clerk? These questions were not treated, for lack of statistical series. Zakhor is in the process of manufacturing the series.

IV. What This Means Politically

Let us say it plainly, for we are asked this question every week.

First front: disinformation. Wikipedia is the reference source that feeds the great artificial intelligence models. On Judaism, on Jewish history, on Israel, its articles are the subject of patient, organized alteration work, and largely won for lack of opposition. Jewish institutions have neither the digital means nor the technical culture to wage that battle.

Our response is not to fight article by article. It is to build, alongside, a corpus so well sourced, so scrupulously qualified, so systematically traceable that it becomes the natural reference of the models themselves. You do not win against falsification by shouting louder. You win by producing something so manifestly superior in sourcing quality that circumventing it becomes costly. This is a wager on rigor. It is the only one that has ever worked.

Second front: identity and resilience. We are launching with the International Institute for Jewish Genealogy a pilot study among American students: a genealogical journey on Zakhor, framed by questionnaires before and after. The hypothesis is simple and, if verified, considerable. A Jewish student confronted with hostility on his campus holds up better when he knows who he is — not in the sense of abstract belonging, but in the very concrete sense of: here is my great-great-grandfather, here is his village, here is his signature. A documented identity resists what an merely asserted identity cannot withstand.

If the pilot confirms the intuition, it becomes a program, and the program becomes a tool deployable on every campus in the Western world. The same demand is rising in France: young generations want to know, and they are saying so.

Third front: archives. An immense part of Jewish memory is held by institutions that preserve it admirably and share it sparingly. Our work in the coming months is as diplomatic as it is technical: to convince, one by one, that openness serves their mission.

V. The Deployment

None of this happens through technique alone. Zakhor advances through human relays: ambassadors by country, by community, by generation, by family. Someone who presents the project at his synagogue one Sunday morning does more for the corpus than three weeks of computing.

This week, Jean-Yves Carfantan — economist, thirty-five years in South America, Jewish by choice, member of Judaism in Motion — has accepted this role and will organize a volunteer meeting at Copernic. Leaflets, talking points, digital tools: everything is available for whoever wishes to do the same. Write to us.

That it should be a man from a Breton Protestant family who carries the project into a community of three thousand families says enough about what this people is, and what this memory is: not a locked genetic heritage, but a book one can ask to belong to.

VI.

We asked last week why memory is a commandment. Here is what five months of work have added to the answer.

Zakhor is an imperative, but it is an imperative that, for three thousand years, could never be kept except in part. One remembered what one could: the great figures, the great texts, the great disasters. The others fell away. Each generation chose, without meaning to, whom it would let slip into oblivion — and the choice was made by default, by lack of space, by lack of hands.

That is what is changing. For the first time, the constraint is no longer material. We no longer have to choose. We can, literally, keep the commandment.

All that remains is to do it. One stone, one act, one name at a time.

In three days, 5787 opens. May it be the year of open archives, of names recovered, of families restored to one another. May it be gentle to each of you, and to those whose names you bear.

שנה טובה ומתוקה
Shana Tova ou'Metouka

Bernard Bensaid
Founder of Zakhor.ai

In this letter. 88 Great Books appeared this week — foremost Flora Sassoon, then 3 lineages, 84 places; 4 doors opened in the collection — Bedarida, Pierre Vidal-Naquet, Congregation Mickve Israel, כתובה. קהיר, מצרים. תרנ"ד. ; one account — The Stele of the God-Fearing ; 315 finds from the day before, of which 3 reported here; what changed on the site, in 7 days of work; 6 markers on the calendar of the eight days to come; 14 arrivals, and here we are 172 keeping this register.

The Grand Book of the Week — Flora Sassoon

There are few figures in the history of Jewish diasporas where world commerce, scrupulous piety and Talmudic erudition meet with such intensity. Flora Sassoon — her birth name was Farha Gubbay

15 pages.

→ Read The Grand Book — Flora Sassoon

What Appeared This Week

88 Great Books were born in seven days. Here they are, by pillar.

The Lineages — 3

Behar — 13 p.
Riedel — 10 p.
Bassoul — 9 p.

The Places — 84

Lincoln — 12 p.
Miami — 11 p.
Millstone — 11 p.
Oungre — 11 p.
Livonia — 11 p.
Windsor — 11 p.
and 78 others — see all

Four Doors into the Collection

A lineage, a figure, a community, a work. Each week, four others.

The Life of a Lineage — Bedarida

There are names that cross borders without ever forgetting them. Bedarida — also written Bédarida, Bedaridda or Bedaride — is one of these. Each of these spellings says something: a language of reception, an administration that transcribed the surname in its own way, a branch that separated from it carrying its own orthography.

→ Read — 14 pages

A Great Figure — Pierre Vidal-Naquet

Pierre Vidal-Naquet (1930–2006) belongs to that rare category of historians whose scholarly work and civic engagement answer each other without ever merging. A leading Hellenist, he renewed our understanding of ancient Greece by interrogating its myths, its margins and its institutions; an uncompromising citizen, he waged three great battles

→ Read — 12 pages

A Community — Congregation Mickve Israel

On July 11, 1733, merely five months after James Oglethorpe landed on the shores of Savannah to found the thirteenth British colony in America, a small ship, the William and Sarah, cast anchor at the same place after a perilous crossing from London. On board were forty-two Jews

→ Read — 13 pages

A Work — כתובה. קהיר, מצרים. תרנ"ד.

The object that opens this volume is a ketubah (כתובה, "document") — a Jewish marriage contract — produced in Cairo, Egypt, during the Hebrew year תרנ"ד, that is the civil year 1893-1894. The simple reading of its title alone suffices to summon a world: that of the Jewish communities of Egypt at the apex of their diversity, at the crossroads of the waning Ottoman Empire, of nascent British guardianship and…

→ Read — 10 pages

Short Stories — The Stele of the God-Fearers

Aphrodisias, Caria (Asia Minor) · 4th–5th century

A marble block nearly three meters high lists the donors to a synagogue in Caria. An entire column is not Jewish — and no one hid it.

Aphrodisias, in Caria, was the city of Aphrodite — its name says so. A marble block nearly three meters high was unearthed there, engraved on two faces, bearing more than one hundred twenty names of donors to a Jewish institution.

→ Read the story · All stories

What the Nights Found

Our watch turns every night, in thirteen languages. 315 findings this week. Here are three.

A magnetic authentication method could expose fakes of Judean antiquities

A team led by Yoav Vaknin, with Jeffrey Gee and Lisa Tauxe (Scripps Institution of Oceanography, UC San Diego) and Brendan Cych (University of Liverpool), published in early September 2026 in the Proceedings of the National Academy of Sciences (PNAS) a method for authenticating ancient ceramics based on archaeomagnetism. Source

"Hungarian Holocaust Revisited": a collective volume takes stock of recent discoveries in Hungarian research

De Gruyter Brill published in June 2026, in its "Holocaust Studies and Antisemitism" series, the collective volume "Hungarian Holocaust Revisited: New Discoveries and Insights" (370 pages, in English), edited by Dóra Pataricza (Országos Rabbiképző – Zsidó Egyetem, Budapest), Borbála Klacsmann (ELTE) and Péter Buchmüller (Central European University). The work presents to the international public the recent advances in Hungarian research on the Shoah, with an emphasis on contemporary approaches: gender studies, microhistory, memory and trauma. Source

New York: the Museum of Jewish Heritage exhibits the clandestine archives of the Warsaw Ghetto (Oyneg Shabes)

The Museum of Jewish Heritage in New York opens on September 18, 2026, until February 14, 2027, an exhibition devoted to the Oyneg Shabes clandestine archives of the Warsaw Ghetto, eighty years after the discovery of the first buried cache in 1946. The exhibition, co-organized with the Emanuel Ringelblum Jewish Historical Institute in Warsaw, traces the secret work conducted between 1940 and 1943 by historian Emanuel Ringelblum and a clandestine group of scholars, educators, journalists and young activists, who gathered nearly 35,000 documents. Source

→ All this week's findings

What Changed on the Site

The cemetery catalog extends to dozens of countries, and the site catches up its translations in all its languages

This week, the catalog of Jewish cemeteries extends to dozens of new countries — from Hungary to Austria, passing through Argentina, Central America, South America, Czechia, Slovakia, Romania, Belarus, Lithuania, the Netherlands, Bulgaria, Latvia and eight Balkan countries — with hundreds of thousands of additional graves linked to their families day after day. A great wave of translation catches up what remained in French: the site's content, the administration console, the error messages sent to visitors and the latest emails addressed to members all move into the site's thirteen languages. The protocol of the campus study on Jewish identity takes its final form — a single group measured twice before the work, then once after — and members can now measure their own Jewish identity from their dashboard. A thorough audit closes several longstanding flaws in the site, including a security breach in the burial register and a loop that made 99% of burial links inaccessible.

Monday, August 31 — British and Ukrainian graves join the catalog, and the debt of place cards is settled
Tuesday, September 1 — The study on Jewish identity gains its journal of gestures, and the linking of graves crosses the alphabet
Wednesday, September 2 — The burial register finally distinguishes visitor from member, and a three-minute leak is sealed
Thursday, September 3 — Hungary, Argentina and Latin America join the catalog, and members can measure their Jewish identity
Friday, September 4 — Seven new cemetery countries, a massive resumption of translations, and the archives of Yad Vashem and Arolsen explored
Saturday, September 5 — A complete site audit closes several longstanding flaws, and Argentina and the Netherlands complete the catalog
Sunday, September 6 — Austria, Bulgaria, Latvia and the Balkans complete the catalog, and a new tool explores kinship links

→ The project, day by day

The Week Ahead

Friday, September 11 (29 Elul 5786) — Eve of Rosh Hashanah
Friday, September 11 (29 Elul 5786) — Inauguration of the Salisbury Road synagogue in Edinburgh — 1932
Saturday, September 12 (1 Tishrei 5787) — Rosh Hashanah
Saturday, September 12 (1 Tishrei 5787) — Death of David Oppenheim, Chief Rabbi of Prague — yahrzeit · 1736
Sunday, September 13 (2 Tishrei 5787) — Rosh Hashanah (2nd day)
Monday, September 14 (3 Tishrei 5787) — Fast of Gedaliah — fast

What You Have Deposited

(no deposit this week — block to be removed, or to be returned as an invitation)

1 796 Published Ledgers · 7 387 lineages · 15 445 figures

Those Who Have Joined Us

14 people opened an account this week. Members reconnected with their family: Bass, Lieberman, Riedel. We are now 172 keeping this register together.

We do not name them: one comes here to search for one's own, not to be announced.

Join in turn →

What You Can Do

Write, enrich, or simply read a Great Book. Every family, every community, every work can have its own. If one already exists, you can add what is missing from it; if it does not exist, you can bring it into being.

→ Write or enrich a Great Book · Browse the Great Books

Bear witness about one of your loved ones. A grandfather, an aunt, a neighbor from the street before. What you know of a person — where they lived, what they did, what they used to say — is written nowhere else. The account is signed with your name and attached to the tree.

→ Testify about a loved one

Import your family tree. If you maintain a tree in software — Heredis, Geneanet, MyHeritage, Gramps — it exports to a GEDCOM file that you can upload here. You remain in control: nothing is published without your action.

→ Import a tree (GEDCOM)

This week, a gesture

Look at what you keep without knowing why. A photo with no caption. A letter in a language you cannot read. A piece of paper at the bottom of a box, that you have never thrown away yet never been able to say why.

Deposit an item →

zakhor.ai

Newsletter

The archive is reserved for members