Book Review: Open Socrates (Part 2)
TL;DR
The speaker thinks Agnes Callard’s entire “paradox” setup collapses once you use probabilistic reasoning — he repeatedly argues that Meno’s paradox, Moore’s paradox, and the supposed truth-vs-error tension are fake problems if you think in Bayesian terms rather than demanding all-or-nothing certainty.
His central indictment of Socrates is strategic manipulation, not inquiry — he frames Socrates as a two-stage operator who first induces aporia by humiliating people, then “midwives” them into new beliefs, comparing the pattern to cult leaders and military break-you-down/build-you-up tactics.
Dialogue helps, but it’s not the essence of thought — rubber-ducking, writing, and conversation can all unlock insight, but he rejects Callard’s stronger claim that thinking is fundamentally social, insisting solitary reflection, self-correction, and internal disagreement are normal and common.
He sees the Socratic demand for single clean definitions as a trap — on questions like justice, virtue, courage, and piety, he says Socrates keeps converting practical, messy questions into impossible definitional absolutes, then “wins” by exploiting ambiguity and contradiction.
The video keeps tying ancient philosophy back to AI alignment — coherent extrapolated volition, formalizing values, and “writing down the good” become the modern stress test for all this, with the speaker warning that even an idealized aggregation of what people would think after reflection could still encode bad models and real value conflicts.
Part 2 ends by teeing up Part 3 as the real contradiction — after spending hours on Socrates’ claim to know nothing, he points to the coming section on politics, love, and death as the moment Socrates suddenly claims special expertise on exactly the topics that matter most.
The Breakdown
The “paradox paradox” opening shot
He starts by flatly rejecting Callard’s framing of Book Two: open-mindedness, inquiry, and truth-seeking do not conceal deep paradoxes so much as philosophers pretending ordinary cognition is mysterious. Meno’s paradox gets the “have you heard the good news of Bayes’s rule?” treatment, and Moore’s paradox is dismissed as a grammar trick plus a bad model of the mind as one unified, perfectly consistent object.
Probability dissolves the fake dilemmas
The speaker’s recurring move is simple: belief is graded, cached, and updated. So “I was 99.9% sure Mommy wasn’t kissing Santa Claus, now I’m 99% sure she was” is not paradoxical; it’s just belief revision. Likewise, the supposed split between believing truths and avoiding falsehoods becomes, in his words, a daily tradeoff between type I and type II error, not some uniquely Socratic puzzle.
Rubber ducking, not mysticism
When Callard says philosophy is less like filling a void and more like untying a knot, he calls that mystical gibberish. He likes the practical observation underneath it, though: talking and writing can surface thoughts you didn’t yet have in explicit form. His gloss is software-engineering plain English — this is rubber ducking — and his objection is to pretending that because dialogue helps generate insight, all genuine thinking must therefore be social.
AI alignment as philosophy’s final exam
Midway through, he zooms out to AI. If we soon have to formalize “the good,” write down what we want, and hand it to systems that act on it, then philosophy stops being a parlor game and becomes engineering with catastrophic downside. He uses Eliezer Yudkowsky’s coherent extrapolated volition as the obvious example: better than one person guessing, sure, but still likely to go badly if people’s underlying models, methods, and values are wrong in systematic rather than canceling ways.
Socrates as gadfly, smartass, and recruiter
The longest stretch is a sustained attack on the heroic picture of Socrates. The speaker says the “I know nothing” pose is tactical low-status camouflage: Socrates baits people into overcommitting, traps them in contradictions, then humiliates them into paralysis. That’s why he reaches for Douglas Adams’ line about respectable physicists finally realizing they can’t stand a smartass, and why he says Socrates is less a truth-seeker than a recruiter trying to convert others into devoting their lives to inquiry.
The false split between destroyer and builder
Callard’s gadfly/midwife distinction gets hammered too. He says refutation and discovery are obviously not the same thing: “light bulb inspector” is not the same role as inventor, even if knowing 1,000 ways not to make a bulb helps. There are real cases where dividing labor between builder and tester is useful, but he thinks Socrates weaponizes the separation to avoid making positive claims while still dominating the frame.
Thinking is not a two-person job
The speaker is especially animated when Callard tries to define thinking itself as social inquiry into untimely questions. That, for him, is the cardinal overreach. He argues that people constantly reassess their own beliefs, suspend judgment, compare hypotheses, notice contradictions, and update alone; saying otherwise is just “interrupting the person doing it.”
Definitions, Meno, and the setup for Part 3
In the final third, he goes after the definitional obsession itself. “What is justice?” and “what is virtue?” are not cleanly answerable the way Socrates demands, and turning them into single-essence questions mostly creates opportunities for rug-pulls. He closes by saying this all sets up the next section neatly: the man who says he knows nothing is about to claim privileged insight into politics, love, and death — which, to him, is where the real contradiction starts.