Some users praised Talmud study as an operating system for reasoning relevant to AI progress, while many others dismissed the claim as outdated for moral alignment, slower than engineering work, or irrelevant to current models.
Based on 4 visible X reactions from 43 accounts; directional sample.
Ask a question below.
Published answers will appear here.
Dean Ball is about to become OpenAI's Head of Strategic Futures. He's joining the company that's being sued for stealing trade secrets from Apple and sued by newspapers for training on copyrighted content. And now he's reading the Talmud for ethical guidance. The timing is either brilliant or tragic — maybe both.
The Talmud is less a book than an operating system for reasoning. It preserves centuries of adversarial debate, recursive argument, minority opinions, and edge-case analysis. Long before peer review or formal logic, it was training people to think in a way that’s surprisingly relevant to AI.
@juddrosenblatt @deanwball Please don’t use the Talmud as a basis for moral alignment, while in some ways it was pretty progressive for its time, there are many things in it which don’t age well
@deanwball policy written about 2023 models while engineers ship 2025 things. the lag is huge. reading the talmud probably is faster.
The scholar compares ancient interpretive law to AI governance frameworks.
there comes a time in every ai policy professional’s life when they realize they have to read the talmud to make further progress, and for me that time was this week
Proud to say I got there decades before anyone else in AI policy. https://twitter.com/deanwball/status/2076029765631484042
oh no please don't let my dad be right about everything https://twitter.com/deanwball/status/2076029765631484042
@deanwball omg @shiraeis 😂 (@deanwball we were recently chatting about this phenomenon)
Yes… https://twitter.com/deanwball/status/2076029765631484042
@deanwball @juddrosenblatt
Some users praised Talmud study as an operating system for reasoning relevant to AI progress, while many others dismissed the claim as outdated for moral alignment, slower than engineering work, or irrelevant to current models.
Based on 4 visible X reactions from 43 accounts; directional sample.
Ask a question below.
Published answers will appear here.
Proud to say I got there decades before anyone else in AI policy. https://twitter.com/deanwball/status/2076029765631484042