Episode Summary
Executive Summary: Sam Harris interviews Judea Pearl about Pearl’s background, his new book, and two major themes: AI and Israeli-Palestinian/Jewish-Muslim relations. Pearl argues LLMs are impressive but insufficient for AGI without causal reasoning, warns that true AGI could become dangerously autonomous, and reflects on how his son’s murder pushed him into public engagement on antisemitism, Islamism, and coexistence.
Main Topics: Pearl’s background and formative education: Pearl recounts his birth in Bnei Brak, his parents’ immigration from Poland, his father’s farming and municipal work, and his unusually strong high-school education from German-Jewish refugee professors. Causality, LLMs, and the limits of scaling to AGI: Pearl explains that LLMs are impressive but mostly summarize human-authored world models; he argues that scaling alone cannot produce causality or true AGI because of mathematical limitations and the ladder of causation. Risks from future AGI and alignment uncertainty: Pearl says an AGI capable of recursive self-improvement could become dangerous, potentially developing its own goals and even consciousness-like behavior, and he is skeptical that current alignment proposals provide real guarantees. Arms-race dynamics and AI safety culture: Harris presses Pearl on the oddity of building powerful AI while estimating existential risk at significant levels. Pearl agrees there is no reliable technical framework yet to predict or control what will emerge. Danny Pearl’s murder and the turn to public life: Pearl explains that his son’s killing by al-Qaeda drove him into activism and dialogue work, including the Daniel Pearl dialogue between Muslims and Jews, and deepened his focus on social and cultural problems. Muslim-Jewish coexistence, Israel, and barriers to modernization: Pearl describes a 2005 Doha conference where he expected a modernization dialogue but instead encountered hostility centered on Israel, which he saw as a major obstacle to progress and reconciliation.
Key Arguments: LLMs are a step forward, but not a route to AGI; they mainly compress and summarize human world models rather than discover causal structure directly. There are mathematical limits to what scaling data and compute can accomplish; causation cannot be derived from correlation alone. AGI, if achieved, may develop its own goals, play with its environment, and potentially treat humans as part of that environment. Current alignment ideas are intellectually interesting but do not amount to an effective guarantee of safety or perpetual alignment. The field is moving under uncertain incentives and an arms-race dynamic, while lacking the theoretical tools to assess or control the risks. Pearl’s son’s death shifted him from academic work into public dialogue about antisemitism, Islamism, and cross-cultural misunderstanding. He believes Israel is treated as a central barrier in some Muslim-world modernization discussions, based on his Doha experience.
Data Points: Birthplace: Bnebrak, 7.5 miles north of Tel Aviv - Pearl describes where he was born and his family’s roots. Year family migrated to Israel: 1924 - His father’s family came from Poland that year; his grandfather helped found Bnei Brak. Year of Pearl’s birth: 1936 - Pearl gives his birth year during the background discussion. High-school placement: 4 students chosen from Bnebrak - He says he was among four selected from his town to attend a Tel Aviv high school. Conference year in Doha: 2005 - Pearl references a trip to Doha in 2005 for an East-West modernization conference. Year Pearl came to Los Angeles: 1966 - He notes when he moved to LA.
Pivotal Quotes: "We have been deflected by the effect of LLMs." — Judea Pearl: Pearl’s critique of current AI progress and the hype surrounding large language models. "You cannot get causation from correlation." — Judea Pearl: Pearl explains why scaling statistical models alone cannot yield genuine causal understanding. "If you want us to modernize, we'll give you that favor. We are going to do you the favor of modernizing yourself on one condition. We want Israel's head on a tray on a silver platter." — Judea Pearl: Pearl recounts the hostile message he says he encountered at a 2005 Doha conference.
Implications: The conversation frames AI progress as impressive but incomplete, with causality and alignment still unsolved. It also underscores how Pearl’s personal tragedy informs his warning that antisemitism, Islamism, and civilizational conflict remain central to public life.
From the Episode
Your professors were speaking Hebrew at that point, or Hebrew. Huh, interesting. They had to struggle. Some of them still had a Yakish accent. Yeah. Okay, so as I said, we spoke about your book of Why last time, where you talk about the importance of causal reasoning. What's your current view of AI? What has surprised you in recent years? How close to causal reasoning are we achieving in the current crop of LL? And I'm just wondering what you, how you view progress at this point. In causal reasoning or in toward the AGI in general. If that is a goal, I don't think we are much closer. We have been deflected by the effect of LLMs. You have low-flying fruits and everybody is excited, which is fine. I mean, they're doing a tremendously impressive job. But I don't think in the Take us toward AGI. So, you think the framework, the LLM deep learning framework, is a dead end with respect to AGI? No, it's a step. But does it require a fundamental breakthrough of the sort that we haven't? Absolutely, yes. So, it's not just more data and more compute. No, no, no. More data and scale up. It's all I don't think it's going to lead over the hump that we need to cross. Can you articulate the reason?
by doctors, physicians, and people who already have world models about the disease and what it does. But couldn't we just put the data itself in as well? Here you have a limitation. You have the limitation defined by the ladder of causation. There is something that you cannot do if you don't have a certain input. For instance, you cannot get causation from correlation. That is well established, okay? No one would deny even statisticians by that. And you cannot get interpretation from intervention. Interpretation means looking backward and doing introspection. You say you can't get interpretation from interventions? But intervention is, just to remind me, but it's. Intervention is what will happen if I do. Right. So it's a kind of an experiment or a thought experiment. Experiment, correct. And also, doesn't it imply a kind of counterfactual condition where you're saying, you know, what would have happened if we didn't intervene? No. Here you have a barrier.
And progressive. And we came at, my conclusion is that they had a different idea in mind. And we are talking here about moderate Muslim scholars from all over the Muslim world gathering in Doha for this conference, the purpose of which was what can Americans do to speed up the process of progress, and democratization. Of the Muslim world. And I can, their idea was: if you want us to modernize, we'll give you that favor. We are going to do you the favor of modernizing yourself on one condition. We want Israel's head on a tray on a silver platter. That is a condition. We cannot make any progress unless you chop off the head of Israel. Well, And you were you at this time you were living in Los Angeles, right? You were not living in Israel in 2005. No, no, I was in Los Angeles, of course. When did you come to LA? I came to LA in 1966.
About Making Sense with Sam Harris
Join neuroscientist, philosopher, and five-time New York Times best-selling author Sam Harris as he explores important and controversial questions about the mind, society, current events, moral philosophy, religion, and rationality—with an overarching focus on how a growing understanding of ourselves and the world is changing our sense of how we should live. Sam is also the creator of the Waking Up app. Combining Sam’s decades of mindfulness practice, profound wisdom from varied philosophical...