The Best Conversations Aren’t the Easiest Ones

In an age of AI companions built to agree with us, an old-fashioned folding table on Israeli streets shows what real conversation is for.

I’ve come to believe something uncomfortable this year, watching how people argue online, in Knesset debates, and increasingly with AI: feeling right and being right are not the same thing. But they feel identical from the inside, and most of us never find out which one we actually are.

Here’s how that happens. An idea can feel completely solid without ever once being tested against real resistance, it only needs to be repeated, often enough, by people you already trust. Hear the same view again and again inside a community that agrees with you, and it starts to feel obviously correct, whether or not anyone has actually pushed on it. The cost of that comfort doesn’t show up inside the community that grants it. It shows up later, when the belief finally meets a moment, a war, a betrayal, a person who won’t back down, that doesn’t care how many people agreed with you.

That’s the real reason to worry about where we’re headed with AI, and it’s not the reason usually given. We keep asking whether AI will replace teachers, therapists, rabbis, lawyers, or writers. I think we’re asking the wrong question. The real one is: what do we believe conversation is for?  and what happens when the easiest conversations never test us?

If conversation is mainly about getting information or feeling understood, AI has a real advantage: no awkward silence, no bruised ego, no apology owed the next day. You close the app unchanged, and it won’t mind. It will also rarely tell you you’re wrong in a way that costs you anything, which lets almost any idea you already hold feel a little more solid, without ever being tested.

But what if conversation is supposed to change us, not just agree with us?

We see less of that nowadays. But thankfully, it’s having a revival on Israeli streets: a folding table and two plastic chairs.

I’m thinking of Matan Yaffe, a reserve officer and social entrepreneur who sets up that table on pedestrian malls across the country, Tel Aviv, the periphery, even Bnei Brak, inviting whoever walks by to sit and argue. Haredim, secular Israelis, settlers, students, retirees. No moderator, no algorithm curating who he talks to next. Just two chairs, a camera, and however long it takes.

What makes it compelling isn’t that people agree by the end, usually they don’t. Nor is it shouting; these aren’t the panel-show pile-ons of cable news or Knesset committee rooms. People are encouraged not to talk over one another. What you see instead is rarer: an idea tested in real time. Someone says what they believe, and instead of a nod or a rebuttal shouted over them, they’re asked why, and then asked again until you can tell, almost in slow motion, whether the idea was built to hold weight or only ever felt solid because no one had pushed on it.

A Haredi man explains, at length, why he doesn’t enlist, and Yaffe pushes back, and the man doesn’t leave. By Yaffe’s own account, one such conversation in Bnei Brak nearly ended in a shove and an overturned table and he chose not to publish that part, because the point was never to collect footage of people humiliating each other.

The most important conversations in life are rarely the easiest ones, the family dinner where people who disagree about the judicial overhaul, or the hostages, or the ultra-Orthodox draft, resist the urge to retreat into silence, or into their phones. These conversations are inefficient, sometimes uncomfortable, occasionally painful. They take a patience that has nothing to do with a server responding in milliseconds, the patience to sit with someone whose views unsettle you, to risk being wrong. They’re also the conversations that build trust.

Israel is as internally divided as it’s been in years, over religion and state, over who serves and who studies, over how the war should end. It cannot out-argue itself through headlines and Knesset shouting matches; it has to sit at the table. Families depend on the same thing at a smaller scale, and so does a beit midrash, where an argument was never about winning but about staying to hear the other side of the page.

A chatbot can help us think, find the right words, rehearse an apology before we walk into the room. But it cannot replace what happens when two imperfect people stay in the room long enough to be changed by one another.  And because it risks nothing, it can never earn the trust a person earns by choosing to stay anyway.

Maybe that’s why a folding table on a street in Zichron Ya’akov or Bnei Brak still draws a crowd, even in an age of infinite digital connection. Something in us still knows the difference between an idea that’s never been tested and one that’s actually earned. The real question was never whether your ideas are right, but whether you’ve ever let them be tested, by someone who wasn’t going to make it easy, and didn’t walk away.