When I'm doing vibe coding or using AI in chat form, I often think of Socrates' maieutics — the Socratic method.
Even when AI seems to give a simple or extremely instant (i.e., immediate) answer, the essence is expanding thought through interaction. So I find myself thinking that human communication may, in fact, be returning to a more essential mode.
Background
An era when information is property. Soon enough, the world overflowed with spam. Everyone, no exceptions, sprays out advertising. They sell goods offline and online. From the home computer to mobile messengers, from shopping malls to banking apps, they advance right up to the front of our daily lives — to our fingertips… no, to our bank accounts.
And as content, information becomes ever more refined, and through the medium of video it even develops into a sort of currency. But as we've already learned from many past experiences, an abundance of channels does not equal smooth communication. For instance, when a road is widened it feels like we'll get home faster, but easier access tends to demand more travel, keeping people busier and longer on the road. So once again today, the "life with an evening in it" gets squeezed.
The internet is similar. Beyond cabling every nook and cranny of the country, wireless communication has now grown as dense as the carbon dioxide in the air. If the internet is just a little faster, it feels like we'll save a little waiting time and finish work a little earlier — but our actual quitting time doesn't follow.
And there's more. This may be a very personal interpretation, but I think social networks and the various IM (Instant Message — KakaoTalk, Telegram) apps, instead of providing smooth communication, manufacture a severance of direct conversation. They produce extremely fine-grained, detailed intergenerational conflicts. They build invisible in-groups, with their own cultures and value systems.
Beyond that, the various "user-optimized" algorithms — YouTube and the rest — keep narrowing the user's world. The most fundamental problem with these algorithms is that what they call "user convenience" is not, literally, convenience for the user's life; it's structured around the targeting convenience of information providers (advertisers). In the neoliberal era, the user is the advertiser's target. The platform binds the user inside its own field of view in order to do that targeting.
We live inside the menu of options the advertiser wants. The choices we have right now are an algorithm distorted by capital. Just as Dragon Ball's instant teleportation has already been realized in another form right in front of us, the world of the Matrix has already been laid out by capital and platforms. The only difference is the visualization style and interface that we used to imagine.
We have to escape the field of view — the so-called frame. To do that, we first have to recognize that we are inside the frame. We then have to develop a critical awareness of the frame. Only then can we have free will about whether to leave or stay. The next requirement is intellectual, physical, and surplus capital. Free will alone won't get you out of the frame. The frame in real life (the routine) needs only will and stamina, but escaping a digital frame — building your own digital frame — costs money. These are genuinely difficult things.
In this process, we end up needing one tool. Namely, the AIs that have recently been launching as a wide range of products. I'll set aside introducing, explaining, or showing how to use those products. That kind of information is everywhere, and people more credible than I am are explaining it more easily.
What I want to say in this long-winded post is the single idea I mentioned at the start. Doing vibe coding or using chat-form AI often feels like Socrates' maieutics. So what other forms might be like maieutics? I jotted down a conversation I had with the AI.
1. Socrates — “The Originator of the Q&A Method”
- Trait: Maieutics — drawing answers out of someone through questions.
- Resemblance to AI: Asking questions and continuing into further questions based on the answers — the same loop as today's conversational AI.
- Motivation point: Using AI not as a mere information source, but as an intellectual partner that elicits thought.
- Form: **Circular**, ring diagram (Q → A → counter-Q → deeper Q … continually expanding).
- Visual: A circle labeled “Pursuit of Truth” at the center, with spiraling Q&A radiating outward.
- Feel: A structure that, like maieutics, goes progressively deeper.
2. Confucius — “Self-cultivation Through Question and Dialogue”
- Trait: Within the brief Q&A records with disciples (the Analects), different answers are given depending on the situation.
- Resemblance to AI: Just as different answers fit different contexts and people, AI tailors responses to the user's context.
- Motivation point: Bringing AI conversations into practical wisdom for daily life.
- Form: **Contextual**, tree/branch diagram.
- Visual: Roots are “the Way of life”, branches are situational questions (filial piety, propriety, the noble person, the petty person), leaves are answers fitted to each.
- Feel: Showing how dialogue flows differently with different contexts.
3. Kierkegaard — “Dialogue That Guides Inner Reflection”
Trait: Emphasizes “existential choice” — writing that poses questions to the reader and lets them wrestle with the answer themselves.
- Resemblance to AI: Helping the user discover the answer within themselves rather than fixing it.
- Motivation point: Treating AI as a mirror for reflecting on one's own existence.
- Form: **Mirror**, symmetrical figure (left: question, right: reflection).
- Visual: Left side “AI's question”, right side “my inner answer”, center marks “decision/choice”.
- Feel: A mirror-like structure where dialogue echoes back from inside.
4. Wittgenstein — “Dialogue as a Language Game”
- Trait: In *Philosophical Investigations*, he develops thought through brief Q&A and examples.
- Resemblance to AI: As meaning is made through “the use of language”, dialogue with AI also expands thought within language games.
- Motivation point: Through AI, we can test the boundaries of language and concepts and explore new meanings.
- Form: **Language game**, lattice or network graph.
- Visual: Word/concept nodes connected by lines, with Q and A moving between different “language games”.
- Feel: Expressing how meaning is shaped not on a single path but across many paths (game rules).
5. Beauvoir & Sartre — “Inter-subjective Dialogue”
- Trait: They unpack freedom and the recognition of the Other through dialogue.
- Resemblance to AI: Although AI has no consciousness, treating it as a “talking Other” expands self-understanding.
- Motivation point: Through AI conversations we can confirm that “the questions I ask reveal who I am”.
- Form: Inter-subjective, two overlapping circles (Venn diagram).
- Visual: One circle is “me”, the other is “AI/the Other”, the overlap is “new understanding”.
- Feel: Dialogue isn't mere information transfer — it's an interaction where “you discover yourself through the other”.
It's starting to feel like a conversation with ChatGPT isn't simply a tool for finding answers. Its interface isn't really different from communicating with a human being. If anything, it can be more friendly and more concrete than person-to-person communication.
My sense is that, today, communication between humans has been heavily distorted in many ways — by what were, in their early days, immature, un-humanistic, neoliberal digital tools. If you hold this kind of life pattern up against the idea that talking with ChatGPT is a sort of maieutics, conversation with ChatGPT ends up looking a lot like a kind of treatment process for the human communication that has been distorted all this time. Of course, that doesn't mean it's a digital "treatment drug"; it's more like a lifestyle or an everyday culture. Back in my day — that is, when KakaoTalk first came out — there was a period when smartphones spread not because of the smartphone's own usefulness, but because of KakaoTalk.
Against that backdrop, my conclusion is that conversation with this kind of AI shouldn't be founded on the prompt engineering techniques the experts talk about, but on the idea I described above. So I'm scribbling down a few rambling, unstructured, will-driven lines for the first time in a while.
— Earlier posts I wrote —
Sayonara, Prompt Engineering
With the ChatGPT craze launched by OpenAI, a newly minted job called "prompt engineer" appeared. Prompt engineering is the practice of guiding generative AI solutions to produce the desired output…
normalstory.tistory.com
