Follow-up on ‘Contemporary Methods of Studying Talmud - Brisk vs. Academia; and the Artifically Intelligent Lamdan’
I received a lot of feedback on my post yesterday:
Thanks to everyone who commented on the Facebook post and reached out (SZ, GS, SL, and MK)!
I wanted to clarify my stance.
Shlomo Zuckier commented:
“Comparing textual versions should be simpler. There are already programs that do that.”
My response: There are specific aspects of the academic method that AI can assist with, true. But my point is that the academic method uses a much wider variety of tools than the reductionist Brisker / analytic / conceptual method. That wider variety of tools likely makes it harder to automate, holistically-speaking.
To clarify further: I have no question that AI will revolutionize academic talmud study. This is something that I'm excited for, and I've been testing it out these past few months.
See also Koppel and Shmidman's really good article on the topic at The Lehrhaus: Torah Study and the Digital Revolution: A Glimpse of the Future” (January 28, 2020).
But my point is that, at least in theory, it is arguably the reductionist conceptual method that is most amenable to LLM AI tools. Though I'll grant that I haven't tested this.
It would require LLM AI training on Talmud Bavli. I know that there's a ton of work and investment happening in the legal industry to incorporate generative AI tools, which is likely methodologically applicable. See especially Harvey, and this Reuters article about it: “Legal AI race draws more investors as law firms line up“ (April 26, 2023) (https://archive.is/sIpBT).
See also my speculations in my recent Seforim Blog post, “From Print to Pixel: Digital Editions of the Talmud Bavli” ( June 5, 2023):
"Given the rapid advancements in artificial intelligence and large language models, it appears highly likely that AI will continue to improve its ability to interpret the Talmud, along with all other sources, including digitized manuscripts. It may not be long before we turn to AI to obtain the ultimate p’shat in the Talmud.[34] Furthermore, there is a possibility that AI could eventually bring comparative sources to bear, as previously mentioned."
Gil Student commented: “Is it possible that this is an area that distinguishes between the methods of Halivni and Shamma Friedman? Meaning, Friedman is more methodical and therefore should be more replicable by another scholar or by AI. Halivni is more intuitive and would be harder for an AI to replicate”
My response: It’s a fair point. But again, similar to what I said to Shlomo, the splitting into layers (revadim) is only one part of it, albeit a central one in contemporary academic method.
Miscellaneous
High-level, simplified taxonomy and charts related to Jewish history, from the Jewish YouTuber “UsefulCharts”:
Jewish Denominations Explained [Reposted with Changes]:
I watched parts, and from the parts I watched, they’re pretty good. I independently made similar charts a few months ago. Of course, there’s plenty of nitpick with, and endless ink has been spilled on all of this. All models are wrong, but some models are useful...