The Shofar: Beyond Words


13 min read
10 min read
5 min read
8 min read
Two Talmudic study partners used to fight over every point for hours. Turns out that's exactly how to get AI to actually teach you something.
We are living in a generation where answers to almost any question are available instantly. When assigned tasks at school or work, we just type in a short prompt to AI and the rest is taken care of.
When you don’t understand the homework, take a photo of it and BOOM, you have the answer with full working in one second.
When you have class in a couple minutes and haven’t read the assigned books, don’t worry. Just type in the title to ChatGPT and problem solved.
When you don’t know what to think, ask AI to formulate your opinion for you.
What is this overused and easily accessible tool doing to us?
In one study conducted in 2025, 580 university students were examined on the cognitive implications of AI on education. The results revealed that greater AI dependence was associated with less critical thinking and greater cognitive fatigue.
Cognitive offloading, using external aids to reduce mental effort and memory demands, may be partially to blame. Cognitive offloading has always been around, from abacuses to paper and notebooks. AI is simply a more powerful version of this same idea.
With AI’s negative impact on education, some are calling to rid ourselves completely of its use. For example, the New York City Public School System has implemented a strict moratorium banning student-facing generative AI tools for pre-K through eighth-grade students.
The problem isn’t that the AI can do these things; it’s that we keep asking it to.
But I wonder if we’re focusing on the wrong culprit. What exactly did AI do wrong?
ChatGPT didn’t just randomly decide one day to write your assignment for you. You asked it to.
AI didn’t decide to complete all your math problems for you. You pasted it in.
The problem isn’t that the AI can do these things; it’s that we keep asking it to.
AI doesn’t need to be a negative burden on our education system. Just like we use notebooks and paper to cognitively offload as a tool to increase our learning, we can do the same with AI. A notebook helps us remember; it doesn’t replace the understanding. AI can be used in the same way, becoming an aid rather than a substitute.
To begin using AI effectively, we need to recognize why it is currently being used ineffectively. Think about your last assignment that you genuinely completed all by yourself. You probably felt stuck, tried something, failed, tried again and maybe again until finally, something you did began to work and you figured it out.
That moment felt different to just having read the solution, because you actually constructed your understanding.
The difficult struggle makes you engage with the problem, making the answer something you genuinely understand.
This is a concept called productive struggle. When a task is difficult, it doesn’t mean it’s an obstacle to your learning. In fact, under the right conditions, that difficulty becomes a key part in how the learning occurs. When you are forced to work on something yourself, make mistakes and attempt different approaches, you are building an understanding rather than just receiving one. The difficult struggle makes you engage with the problem, making the answer something you genuinely understand.
Jewish learning places remarkable value on productive struggle. Think about traditional Jewish study; two Jews sit opposite each other with a text, one thinks one thing and the other disagrees. They argue over it and explain their reasoning, check other sources, argue some more, keep trying and eventually come to a conclusion.
An outsider may look at this and think, “That seems extremely inefficient, why don’t they just ask a rabbi? Or someone who knows the answer?”
There is a meaningful story of two rabbis that emulate this concept, a story about Rabbi Yochanan and Reish Lakish (Bava Metzia 84a). These two rabbis were partners in learning, until Reish Lakish passed away. Rabbi Yochanan was given a new partner, and when he made an argument, his partner brought him evidence showing that he was correct.
Rabbi Yochanan didn’t need someone to prove that he was right; he needed someone to challenge him, forcing him to prove why he was right.
Rabbi Yochanan was not happy. Not because his new partner was unintelligent, but because he didn’t challenge him. When he and Reish Lakish would learn together, Reish Lakish would attack his arguments with 24 objections, and Rabbi Yochanan would respond with 24 answers, repeated back and forth until the subject became clearer.
Rabbi Yochanan didn’t need someone to prove that he was right; he needed someone to challenge him, forcing him to prove why he was right.
Effective learning involves many steps:
Question -> attempt -> challenge -> failure -> reconsideration -> understanding -> answer
However, we often use AI like this:
Question -> answer
Instead of removing AI, what would happen if we apply these same principles to AI?
AI can: challenge you, understand and adapt to your level, generate objections, give hints, ask questions, identify misunderstandings, quiz you, test you and so much more.
And it can do it at any time, with extremely quick responses.
I am not saying we should use AI to replace teaching, however, I do believe it can make a positive impact on the education system itself. There is strong evidence backing this up too.
In a 2025 randomized trial involving 194 Harvard undergraduate physics students, researchers compared learning in an active classroom lesson with an AI tutor designed using established educational principles.
The results were striking: students using the AI tutor showed higher learning gains in less time, as well as greater engagement and motivation.
How can you apply this to your own learning?
Instead of just getting AI to solve your problems for you, ask it to question and support you until you solve it yourself.
You can ask for hints when you are stuck and ask it to help guide you with issues that come along the way. Instead of getting AI to explain a concept for you, try explaining what you know of it and let AI identify the gaps.
Rather than getting it to summarize a chapter for you, read it and get AI to quiz you on it.
Instead of getting it to write an argument for you, write your argument and get it to challenge all the weak assumptions and points you make.
Let AI become your Reish Lakish.
The truth about whether AI strengthens or weakens your learning ultimately comes down to you. AI amplifies how you choose to learn.
Use it to escape effort, and it will hold you back. Use it to challenge and uplift you, and it will become an incredible personalized tutor.
Thousands of years ago, Jewish learning built itself around an idea we are now being forced to rediscover. The quickest path to an answer is not always the best path to understanding.
Don’t just chase after the solution, appreciate the process and allow yourself to reach the conclusion.
Let’s stop blaming AI and instead recognize our own habits and start using AI as a tool, not an escape.

This is such an interesting outlook at AI and a very thought provoking read. I will never use AI the same way again!
Great insight