You Cannot Retrain a Two-Thousand-Year-Old Model
Three years after October 7, the Jewish people continue to confront an unrelenting wave of vilification, open hostility, and blood libels. What is perhaps most striking is not only the intensity of this hatred but also its timing. Demonstrations erupted across university campuses, city squares, and international institutions while Israelis were still burying their dead, before the IDF had fully launched its military response against Hamas.
That chronology matters. Accusations of genocide, colonialism, and illegitimacy cannot be explained simply as disagreement with Israeli policy. They reflect something deeper – a framework in which Jews and the Jewish state are judged before the evidence is even examined. Most recently, these accusations have become political slogans repeated by the current mayor of New York City, the U.S. senator-elect from Michigan, and a former talk-show host who has rebuilt his public identity around anti-Israel hostility. Adding to this dynamic, anti-Zionist organizations such as JVP and IfNotNow, which present themselves as Jewish voices of opposition to Israel, along with a loud chorus of individuals of “Jewish extraction” (to borrow a phrase from Farewell to the French Republic), have lent a veneer of Jewish legitimacy to narratives fueling the broader campaign of hatred.
Many argue that Israel is losing the propaganda war. The proposed solution is familiar – better messaging, more funding, stronger spokespeople, more sophisticated social-media campaigns, perhaps even AI-powered public diplomacy. Others believe that replacing Israel’s current democratically elected government with another would somehow fundamentally change the world’s perception of the Jewish state. It is remarkable how many people continue to believe this nonfactual nonsense, despite the long record of Jewish history and very recent experience.
The problem is that this diagnosis misunderstands the challenge itself. Like every democracy at war, Israel deserves scrutiny and criticism. But the casual invocation of genocide demonstrates how easily a political narrative can overwhelm both historical memory and legal meaning. The term has increasingly become a catch-all accusation, often detached from the realities of how the war is actually conducted. Whatever criticism one directs at Israel’s military campaign, any serious judgment must also account for the extraordinary measures the IDF has taken to reduce civilian harm, including advance warnings, evacuation efforts, and humanitarian measures. Those facts cannot simply be discarded because they complicate a preferred narrative.
Genocide is not merely a synonym for a war or even for allegations of unlawful military conduct. The term was coined by Raphael Lemkin, the Polish Jewish jurist whose work gave legal language to the deliberate destruction of a people. Lemkin escaped. Three million Polish Jews did not. Words carrying that historical weight should not be reduced to political slogans simply because they can be applied to something one opposes.
The same pattern appears in the increasingly fashionable description of Zionism as a colonial enterprise. That narrative begins in the middle of the story rather than at its beginning, a deliberate choice that allows thousands of years of Jewish history to disappear. A former U.S. president employed precisely this framing in an address in Ireland, presenting the modern Jewish return through the familiar vocabulary of anti-colonialism. Once the starting point is chosen conveniently enough, almost any historical narrative can be manufactured.
But Jews are called Jews because of Judea. Their religious, cultural, and political identity originated in the Land of Israel long before Christianity or Islam emerged. Modern Zionism did not invent this connection. It translated an ancient attachment into practical reality through immigration, land purchases, institution-building, and the revival of Hebrew. In many ways, it represents one of the largest decolonization movements in modern history.
Facts matter. Communication matters. Public diplomacy matters. Appealing to the wisdom and moral judgment of those around us matters as well. But history teaches that deeply rooted prejudice is rarely defeated by facts or messaging, however sophisticated or persuasive. Were the Jewish communities of the Rhineland massacred during the Crusades because they lacked an effective spokesperson to explain their case to the armies marching toward the Holy Land? Perhaps Isaac Abravanel and Luis de Santángel were insufficiently passionate, persuasive, or well-funded to convince the Spanish monarchs to revoke the Alhambra Decree? Or should we believe that the hundreds of thousands of Jews slaughtered by the Cossacks during the Khmelnitsky massacres simply failed to explain their place in Polish society clearly enough to their killers? Had German Jews failed to demonstrate their patriotism and deep identification with Germany? Did Stalin’s Jewish Anti-Fascist Committee perish because it failed to explain its contribution to the Soviet victory over Nazi Germany? Or were the Jews murdered in the Farhud simply not integrated enough into Iraqi society to prevent the pogrom?
History answers every one of these questions with the same devastating clarity – the problem was never inadequate Jewish messaging, insufficient loyalty, a lack of public relations, or a failure to belong. These tragedies unfolded under different rulers, religions, ideologies, and historical circumstances. Yet the underlying prejudice proved remarkably persistent.
Borrowing a metaphor from artificial intelligence, antisemitism is a mental model trained over more than two thousand years. Its training set contains religious polemics, conspiracy theories, racial doctrines, nationalism, communism, fascism, post-colonial ideology, and now the algorithmic amplification of social media. Its features evolve, but its output remains remarkably stable.
Two thousand years of Jewish exile, statelessness, and vulnerability continuously supplied this well-trained model with new inputs. Ironically, one of its recurring outputs was the demand that Jews leave the societies in which they lived and return to Palestine, the land from which Jewish history, identity, religion, and peoplehood originated. Then, in May 1948, something extraordinary happened – the model received the very output it had demanded. The Jews returned to sovereignty. Suddenly, the neural model lost one of its oldest and most powerful inputs – the image of the stateless Jewish wanderer, dependent on the tolerance of others. But that input did not remain absent long enough for the model to genuinely retrain. Soviet and KGB strategists proved extraordinarily adept at supplying a replacement narrative. The diverse Arab hamulas of Gaza and Hebron, together with the Arab population of the West Bank that had lived under Jordanian rule, were increasingly brought together within the emerging framework of Palestinian peoplehood, while the conflict itself was reframed through the powerful vocabulary of anticolonialism.
The inversion was extraordinary. The Jewish people, returning to the land at the center of their history after two thousand years of exile, could now be portrayed as foreign colonizers. The model simply reweighted itself around the new feature and returned to its familiar equilibrium. The feature had changed. The inference outcome had not.
Of course, the model has never been universal. Deep and genuine support for Israel and the Jewish people among millions of Christians in North America and around the world, as well as within moderate Muslim societies, has challenged the model’s dominance and prevented it from operating uncontested. But the years since October 7 have supplied the model with an enormous new volume of training data. The object of hostility is now a Jewish state roughly the size of New Jersey, one that has continued to demonstrate remarkable economic and technological resilience even while fighting a multi-front war. It was revealing to see, in the months following October 7, how many Jews appeared to expect that decades of enthusiastic support for progressive and social-liberation movements would translate into at least some measure of sympathy, if not open support, for the Jewish people at a moment of extraordinary trauma. That expectation reflected a fundamental misreading of the model. Thousands of years of deeply embedded training were not going to be undone, or even substantially reweighted, by several decades of Jewish participation in the progressive causes of the day. The model had absorbed those alliances without fundamentally changing its underlying inference.
There is another useful parallel from machine learning. A model trained on an enormous dataset does not fundamentally change because we feed it a few new examples. Fine-tuning can change some responses, but the deeply learned representation is much harder to move. The same is true here. Better messaging can persuade individuals and influence particular audiences. What it cannot reasonably be expected to do is retrain, within a few years, a mental model whose prejudices have been reinforced for more than two millennia.
The Torah portions describing Abraham’s journey, which we read during these weeks, offer a different perspective. Abraham set out to embrace monotheism and introduce a new moral vision into a world dominated by idolatry. He did not wait for universal approval before acting. He left the familiar, forged alliances, fought when necessary, pursued justice, and built. In one of his most consequential acts, Abraham purchased the Cave of Machpelah. He did not merely declare his connection to the land. He transformed it into an enduring reality. Thousands of years later, modern Zionism followed a remarkably similar path. Zionists did not wait for the world to accept their historical argument. They built communities, purchased and cultivated land, revived Hebrew, established institutions, and ultimately restored Jewish sovereignty.
Abraham’s answer was not persuasion alone. It was construction. Zionism’s answer was much the same. Following in Abraham’s footsteps means focusing our energy on our own civilizational mission rather than spending our lives trying to retrain the models operating in the minds of millions around the world. This does not mean that public diplomacy is irrelevant. Facts matter. Falsehoods should be challenged. Historical truth should be defended. But these are tools, not the strategy itself.
The enduring Jewish strategy has always been deeper – strengthening Jewish education, building resilient institutions, settling Judea and Samaria, the cradle of Jewish civilization, fostering economic and technological excellence, defending Jewish sovereignty, and acting according to enduring moral principles regardless of how popular those principles may be.
Trying to retrain a two-thousand-year-old model is largely a futile endeavor, particularly in an age when AI systems and algorithm-driven platforms themselves learn from, reproduce, and amplify many of the same historical narratives and social patterns across millions of users. Sometimes the wiser strategy is not to fine-tune the existing model or devise a better reward function. It is to introduce another force into the system. That was Abraham’s genius. He did not spend his life trying to retrain the civilization around him. He introduced a new vector and followed it. Thousands of years later, that vector is still moving.
Perhaps our task is not to persuade the world to redraw its model of the Jews, but to continue building a Jewish civilization strong enough, moral enough, and confident enough to change the world around it. The answer is to build, defend, create, and teach. Like Abraham, we continue moving in the direction we know to be right without making the world’s approval a prerequisite for taking the next step.
Abraham established a new civilizational reference point. Our responsibility is to continue from there. That is not propaganda. It is the Jewish strategy.