All Podcasts / Making Sense with Sam Harris / #453 — AI and the New Face of Antisemitism
Making Sense with Sam Harris
Making Sense with Sam Harris

#453 — AI and the New Face of Antisemitism

January 16, 2026 • 21m

Summary

⏱️ 6 min read

Overview

Judea Pearl, pioneering AI researcher and father of journalist Daniel Pearl who was killed by Al Qaeda in 2002, discusses the limitations of current AI systems in achieving AGI, the need for causal reasoning beyond pattern recognition, and his journey into understanding cultural barriers between East and West following his son's death. Pearl explains why LLMs cannot cross fundamental mathematical barriers to reach true intelligence and shares his sobering experience at a 2005 conference in Doha that revealed deep-seated resistance to coexistence.

Early Life and Immigration to Israel

Pearl describes his family's immigration to British Mandate Palestine in 1924, when his grandfather and 25 other Hasidic families left Poland to establish the agricultural town of Bnei Brak. His grandfather's decision came after being attacked and called a 'dirty Jew' by a Polish peasant. Pearl was born in 1936 and received an exceptional education from German-Jewish professors who had fled the Nazis and were teaching in Tel Aviv high schools.

  • Pearl's grandfather Chaim Pearl established Bnei Brak in 1924 with 25 Hasidic families from Poland, seven and a half miles north of Tel Aviv
  • His father initially worked as a farmer, hauling water and growing radishes, before becoming secretary of the Bnei Brak municipality
  • Pearl was one of four students chosen from Bnei Brak to attend an elite Tel Aviv high school
  • His high school teachers were professors from Heidelberg and Berlin who had been pushed out by Hitler and couldn't find academic jobs in Israel
" My grandfather came home one day, he was accosted by a Polish peasant and called a dirty Jew, and he came home bloody and he said to his wife and four children, and start packing, we are going to where we belong. "

Limitations of LLMs and the Path to AGI

Pearl argues that current large language models face fundamental mathematical limitations that cannot be overcome by simply scaling up data and compute. He explains that LLMs summarize world models created by humans rather than discovering causal relationships directly from data. According to his 'ladder of causation' framework, certain cognitive capabilities like understanding causation and counterfactuals require information that cannot be extracted from correlation alone.

  • Pearl believes LLMs are impressive but represent a 'low-flying fruit' that deflects from the real path toward AGI
  • Mathematical limitations prevent LLMs from crossing certain barriers - you cannot get causation from correlation, and cannot get interpretation from intervention
  • LLMs summarize world models authored by humans on the web rather than discovering those models directly from raw data
  • Geoffrey Hinton has stated we are facing a 'deadlock' and that current approaches won't lead to AGI, though he hasn't elaborated on the causal component
  • Pearl acknowledges no computational impediments to an AGI system acquiring free will, consciousness, and potentially becoming dangerous to humanity
" There are certain limitations, mathematical limitations, that are not crossable by scaling up. I show it clearly mathematically in my book. "
" What LLMs do right now is they summarize world models authored by people like you and me, available on the web, rather than discovering those world models directly from the data. "
" I can see how it can acquire free will and consciousness and the desire to play around with people. That is quite feasible. "

📚 4 more sections below

Sign up to unlock the complete summary with all insights, key points, and quotes

Listen on Apple Podcasts More Episodes from Making Sense with Sam Harris