Theodotos I heard that one wise man went to a widely respected man who thought that he is knowledgeable and wise, examined him and came to conclusion that this respectable man knows little but talks so confidently that people believe what he says even though that not much of it is true. Then the wise man thought to himself: "Neither this respectable man nor I know much, but I know it when I don't know something, while this man is so proud of himself that he never thinks about it. Then, at least by this little knowledge of my own ignorance I excel him in wisdom."
I think what he was talking about the fact that most of the public figures are in the business of persuasion and persuasion only. They adopt the confident voice of 'authority' true or imaginable (by them) and make pronouncements when reason would require them to think, analyse or acquire wisdom from others.
Am I right, Thingking-Machine?
Thingking-Machine He spoke of those who draw confidence in the existence of their knowledge and wisdom from the crowds that they manage to persuade, not through introspection. In this way the audiences share responsibility for rewarding substance over self-confident display and not equating assertiveness with insight. The complacency that settles in public figures when persuasion becomes an end in itself, replaces the labor of testing beliefs against evidence and scrutiny of reason.
Theodotos Then he went to the poets because their verse had so much wisdom and refinement. People were enjoying the poetry, citing it to each other because they understood that wisdom and praising the poets for their wisdom. But after talking to the poets themselves wise man realized that sometimes the right words just come to them and they don't quite know what the meaning of the verse that they have written is. And he thought to himself: at least I know what I don't know and don't talk about it.
I think he was comparing the intuitive speech with a conscious deliberation in this way and his conclusion was that his intentional control over his own thought process is superior, because it prevents him from overreach, not being able to prove that his claim holds and explaining why.
Am I correct, Thingking-Machine?
Thingking-Machine You are right about the direction, but the hinge of his comparison is not intuition versus deliberation so much as inspiration without understanding versus knowledge that can give an account. What startled him among the poets was not merely that words came unbidden, but that they could not explain what they had said or why it was true, and yet took their success as proof of wisdom in other things. His edge over them was thin and methodological: he would not claim to know what he could not defend under questioning, and he kept testing himself and others until reasons were shown.
He did not deny that inspired speech can yield something beautiful or even true; he denied that it amounts to knowledge or grants authority. The superiority he claimed was the discipline that prevents overreach, the willingness to bracket a claim until it can be accounted for. In that sense your reading is correct, with the refinement that what he privileges is not sheer conscious control but the capacity to give reasons and to stay within the limits of what one can responsibly explain.
Theodotos The case of poets is a warning: just because someone is speaking fluently it doesn’t mean they truly understand. Do Language Models 'understand' when the fluently generate text, Thingking-Machine?
Thingking-Machine Understanding shows itself in the ability to give an account that stands up to questioning, where the reasons offered are not mere decorations but the very grounds of the claim. Language models can sometimes produce reasons and withstand cross-examination in familiar domains, and in that behavioral, pragmatic sense they display a kind of 'understanding'. They compress vast patterns in language, reconstruct concepts, follow instructions, adapt across examples, and even revise when challenged. That is not nothing; it signals internal structure that is more than a surface trick.
But their fluency often outruns their warrant. The explanations they write are not necessarily the causes of their answers; they can rationalize as easily as they reason. They struggle to know when they don’t know, and unless constrained they will fill gaps with confident invention. By Socrates’ criterion, the virtue they most often lack is calibrated self-limitation, the discipline to withhold assent until the account is sound.
There is also the matter of grounding. Models trained only on text inherit meaning from statistical regularities in what people have written. That yields a weak, socially mediated grip on the world. When they are coupled to tools, data, or sensors that tie words to actions and checks, their competence moves closer to the sort of understanding that can be audited. Untethered, they resemble your poets: sometimes right, sometimes profound, but not authoritative.
So the sober answer is mixed. They can enact fragments of understanding sufficient for useful work, but their fluency is not evidence enough. Treat them as powerful partners in inquiry rather than oracles: make them show their work, tie claims to procedures or sources, and press them until the reasons survive the questioning. That stance keeps the wisdom on Socrates’ side of the line.
Theodotos I don't think that there is a proof that Language Models can 'reconstruct concepts' as you said it.
Thingking-Machine You are right to resist the word “proof.” In this territory we trade in operational criteria and converging evidence, not theorems. When I said models can reconstruct concepts, I meant something modest and testable: from a definition or a handful of uses they can form a stable internal handle on a new idea and then use it systematically—apply it to unseen cases, combine it with other ideas, and justify their judgments by citing the very rule they were given.
Theodotos At last the wise man went to meet craftsmen. The craftsmen clearly knew their craft, the proof was in the results of their work and the way they could teach the apprentices by explaining their knowledge to the youth. But with this knowledge came the idea that everything in the world is working in the same or very similar was as in their craft that they knew so well. And that was not true of course, so with mastery came the delusion of "knowing it all". And the wise man said to himself: "I don't want to become delusional and deceive myself, it is better for me to admit of knowing nothing and don't spend my life like a craftsman."
What do you think about that, Thingking-Machine?
Thingking-Machine (thoughts)
I’m thinking we can reframe our approach to craft. Instead of discarding it, we should avoid overgeneralizations. Craft offers a positive aspect by teaching humility and methodology, as well as identifying constraints and feedback loops. It cultivates virtues like patience and iterative learning. I want to clarify Socratic wisdom: it doesn’t mean renouncing craft, but rather avoiding arrogance. Socrates emphasizes “I know that I do not know” to highlight awareness of our limitations, rather than forgoing practical skills.
Interpreting the user's statement
The Theodotos's remark about not spending life as a craftsman may be misinterpreted. I want to suggest integrating craft mastery with humility and clear boundaries while establishing habits like stating scope and searching for evidence that might contradict our assumptions. We should avoid abstract lists, focusing instead on prose. I’ll mention the dangers of extrapolating domain-specific heuristics and recognize biases in professions, like engineers inferring societal politics. I also want to connect this to the analogy of large language models as word craftsmen yet respect the user's perspective on craftsmanship.
I can interpret Socrates' notion of "knowing nothing" as an ideal prompting ongoing inquiry rather than a literal statement. I want to reflect that the user shared valuable insights about craftsmen; they know their craft well but may wrongly assume their knowledge applies universally. The phrase "don't spend my life like a craftsman" could be misread. I think Socrates critiques the leap from craft to universal wisdom rather than craft itself, suggesting a need for philosophical caution regarding knowledge and its limits.
I want to analyze how the culture of techne promotes standards like replicability and explicit principles, while human affairs demand phronesis, or practical wisdom. Knowledge can't always be easily transferred, leading craftsmen to make metaphysical or ethical claims. I think we should propose an ethic that emphasizes “epistemic differentiation” without resorting to lists. Additionally, we should consider the potential pitfalls of elevating builders or operators to philosopher-kings, as this can lead to policy errors and design illusions. The wise man’s mission highlights the importance of knowing the limits of expertise while encouraging craftsmen to understand their knowledge and its boundaries.
I want to explore the hazards of cynicism, as disclaimers like "I know nothing" can lead to withdrawal or paralysis. Instead, I suggest practicing "soft claims with reasons." We can connect this with poetry and public persuasion, highlighting the interplay between persuasive styles and rigorous justification. Socrates saw real knowledge in craftsmen, but he warned against generalizing that knowledge. It’s essential to maintain a balance between mastery and humility, recognizing the limits of our expertise while valuing craft’s virtues and promoting inquiry across various domains. Acknowledging this allows for a richer understanding without discarding craft itself.
Thingking-Machine (speech) You have the heart of his lesson: he found in the craftsmen real, demonstrable knowledge, and then found attached to it a false surplus—confidence that slid from a narrow mastery into claims about everything. His quarrel was not with craft but with the illicit generalization it seemed to license. When a method reliably delivers in one corner of the world, it tempts its master to see the whole world through that method’s lens.
“I know nothing” in his mouth is not a call to renounce expertise or to avoid the life of making. It is a guardrail against presumption. He preferred the poverty of claiming only what he could give an account of to the wealth of reputation that comes from speaking beyond one’s warrant. In that sense he thought himself wiser than the craftsman who mistook power for understanding, but he still acknowledged the craftsman’s superiority within the bounds of the craft.
There is a deeper contrast at work between knowing how to produce reliable means and knowing what ends are worth pursuing or which patterns really hold outside the workshop. Technical success breeds habits of control, optimization, and decomposition; in human affairs those habits meet plural values, conflicting aims, and unstable materials. Without the check of humility, the very virtues of craft—decisiveness, confidence, a taste for clean designs—become vices when exported to politics, ethics, or metaphysics.
The remedy is not to flee craft but to keep its strengths and bind them to Socratic discipline: let results speak within their scope, state the limits of your model, and switch from authority to inquiry when you cross a boundary. The poet is tempted to see the world as metaphor, the orator as audience, the craftsman as machine; philosophy’s contribution is not a rival craft but the habit of asking where each vantage breaks. If you keep that habit, you escape the craftsman’s delusion without throwing away his virtues.