필사 모드: The Complete Philosophy Guide for Engineers: Epistemology, Ethics, Free Will, Stoicism, Buddhism, Popper, Pragmatism, and Tech Ethics (2025~2026)

"The unexamined life is not worth living." — Socrates

Engineering is a job where you meet philosophical questions every day:

• Is this data true? (epistemology)
• Is this feature right? (ethics)
• Do users really choose freely? (free will)
• What matters most in a life? (axiology)

Most engineers dodge these questions, or answer them on intuition. The result: good technology turned into bad products. Cambridge Analytica, the harm done by the Facebook algorithm, AI bias — all of it came out of an absence of philosophical thinking.

This post translates the core of 2,500 years of philosophy into an engineer's language. Not as general culture, but as a thinking tool you use every day.

1. Why an Engineer Needs Philosophy

1.1 Technology Is Not Value-Neutral

"We shape our tools, and thereafter our tools shape us." — John Culkin (reading McLuhan)

• Facebook Newsfeed: started out looking like neutral technology → amplified outrage and polarization.
• Uber Surge Pricing: an efficiency algorithm → an ethical problem during a disaster.
• YouTube Recommendation: optimizing for watch time → extreme content spread.

A technical decision is a philosophical decision. It is a decision about which value comes first.

1.2 Signs of Philosophical Deficiency in an Engineer

• "This is just implementation" (dodging ethics).
• "The data speaks for itself" (immature epistemology).
• "Good intentions are enough" (ignoring consequences).
• "The company decided" (dodging moral responsibility).

1.3 Three Ways to Learn Philosophy

1. Historical: the philosophers in chronological order.
2. Topical: by subject (ethics, epistemology, metaphysics).
3. Applied: problem-first (tech ethics, AI ethics).

Recommended for engineers: Topical + Applied. Higher return on time.

2. Epistemology — How Do We Know

2.1 The Definition of Knowledge (Plato)

"Justified True Belief"

Knowledge = truth + belief + justification. All three are required.

In engineering practice:

• What you saw in the data: the belief.
• Repeated verification: the justification.
• Comparison against a control: approaching the truth.

2.2 Cartesian Doubt

"Cogito, ergo sum." (I think, therefore I am.)

Doubt everything, and take only what cannot be doubted as your starting point.

Applied to engineering: training yourself to doubt the foundational assumptions. Do not assume "this API always works" — actually test it.

2.3 Hume and the Problem of Induction

"A thousand sunrises do not guarantee tomorrow's sunrise."

All empirical knowledge is only probabilistic induction.

Applied to engineering:

• "It ran in QA, so it will run in prod" has no grounding.
• Sample size, edge cases, load, and time are all different variables.
• Examples: Y2K, the leap second bugs.

2.4 Popper and Falsifiability

"Good theories are testable (falsifiable)."

The mark of real science is not verifiability but falsifiability.

• You have to be able to say "this model will be wrong in case X."
• Unfalsifiable = not science (astrology, conspiracy theory).

Applied to engineering:

• Hypothesis Testing: "if X, then it must be Y."
• Unit Test = Falsifiability: the case that can come out wrong.
• AB Test: the possibility of rejecting the null hypothesis.

2.5 Bayesian Epistemology

• Prior: the belief you held beforehand.
• Likelihood: the relative probability of the new evidence.
• Posterior: the updated belief.

All knowledge is probability, and it has to be updated.

Applied to engineering:

• Update on the evidence: no digging in.
• Strong claims need strong evidence: Sagan's Razor.
• Guard against overconfidence: measure forecast accuracy with a Brier score.

3. Ethics — What Is Right

3.1 The Three Great Ethical Traditions

1. Consequentialism — Utilitarianism:

   • "The greatest happiness of the greatest number" (Bentham, Mill).
   • Judged by the consequences of the act.
   • Engineer-friendly (it is measurable).

2. Deontology:

   • "Some acts are right or wrong regardless of consequences" (Kant).
   • Universal law, duty, rights.
   • The foundation of privacy and consent.

3. Virtue Ethics:

   • "What kind of person should I become" (Aristotle).
   • Habit and character.
   • The foundation of craft and integrity.

3.2 Comparing the Three Frameworks

In engineering practice: apply all three frames to every situation and check whether the conclusions converge.

3.3 The Trolley Problem — A Hard Choice

A trolley is heading for five people, and pulling the lever redirects it so that only one dies.

• Utilitarian: pull it (5 vs 1).
• Kantian: a person must not be used as a means (pulling = killing one person).
• Virtue: the intuition of an ordinary decent human being.

Applied: the algorithms in self-driving cars. The MIT Moral Machine project.

3.4 The Moral Responsibility of an Engineer

• "I only built the tool" → the Eichmann defense, and ethically insufficient.
• Foreseeable Harm: you are responsible for harm you could have foreseen.
• Dual Use: recognize the gap between intent and actual use.

Engineering examples: facial recognition, deepfakes, surveillance, emotion-manipulating ML.

4. Free Will and Determinism

4.1 Hard Determinism

• Every event is a product of causation.
• The brain obeys physical law.
• Free will is an illusion (Sam Harris).

4.2 Libertarianism (Free Will Is Real)

• If there is to be moral responsibility, free will is required.
• Does the indeterminism of quantum mechanics leave room for it?

4.3 Compatibilism

• Even if determinism is true, free will is still real.
• "A choice that was not coerced" is what free will means.

4.4 Applied to Engineering

• The illusion of user free choice: behavior steered by defaults, recommendations, and notifications.
• The question of responsibility: who is responsible when a self-driving car crashes?
• The ethics of the nudge: Thaler's nudge vs. the dark pattern.

The claim that "the user chose it themselves" is often a result of the design.

5. Stoicism, Revisited More Deeply

5.1 A Quick Review of the Basics

(Covered in the earlier post, "A Meaningful Life.")

• Distinguishing what is in your control from what is not.
• Virtue (wisdom, courage, justice, temperance).
• Amor fati.

5.2 Marcus Aurelius and "Meditations," Translated for Engineers

"You have power over your mind — not outside events. Realize this, and you will find strength."

For an engineer: a failed deploy, a performance rating, a rejected promotion are all outside. Your own reaction is what you control.

"Waste no more time arguing what a good man should be. Be one."

For an engineer: rather than debating "what a good engineer is," behave like a good engineer today.

5.3 The Limits of Stoicism

• Quietist passivity?: acceptance taken too far weakens the drive to change anything.
• Emotional suppression?: the worry that healthy expression of feeling gets repressed.
• Community?: criticized for an individualist bias.

The fix: combine Stoicism with virtue ethics. Personal equanimity plus contribution to a community.

6. Buddhism and Software

6.1 The Four Noble Truths

1. Dukkha (suffering): there is suffering in life.
2. Samudaya (origin): suffering comes from attachment.
3. Nirodha (cessation): put attachment down and suffering ends.
4. Magga (path): the way is the Noble Eightfold Path.

6.2 Impermanence (anicca)

• Everything changes.
• Attachment is the cause of suffering.

Translated for engineers:

• Frameworks, technologies, and companies all change.
• Never pin your identity to one particular tool.
• "Without X I am finished" is an illusion.

6.3 Non-Attachment

• Not "do not enjoy it," but do not try to own it.
• Concentrate on the process without clinging to the result.

In engineering practice:

• The code I wrote is not mine. Criticism in a PR becomes easy to take.
• Do not hang your identity on whether a promotion comes through.
• Do your best, and let go of the outcome.

6.4 Zen and Programming

• Beginner's Mind: shoshin (from the earlier post).
• Presence: be entirely here, in this piece of code.
• Simplicity: the beauty of Zen connects to the beauty of Rams.
• Humor: the humor of a Zen master, the dad joke of a programmer.

7. Pragmatism — The American Philosophy

7.1 William James and John Dewey

• James: "the true is what works."
• Dewey: centered on experiment, experience, and improvement.

7.2 Engineers Are Pragmatists

• Works in practice, not just theory.
• Iterate and learn.
• No perfect plan, just good enough to ship.
• Tools are evaluated by results.

7.3 The Trap in Pragmatism

• Focusing only on short-term utility → long-term principles get eroded.
• "It works, so it is fine" → ethics ignored.

The fix: pragmatism plus principles.

8. Tech Ethics — Practice in the AI Era

8.1 Data Ethics

• Consent: is it really informed consent.
• Purpose Limitation: no use beyond the original purpose.
• Data Minimization: collect the minimum needed.
• Right to Be Forgotten: GDPR.

8.2 Algorithmic Bias

• Historical Bias: learning from data that carries past prejudice.
• Representation Bias: a group under- or over-represented.
• Measurement Bias: bias in the proxy metric.
• Aggregation Bias: subgroups ignored.

In practice:

• Demographic analysis of the training data.
• Fairness metrics on the output (demographic parity, equal opportunity).
• Shadow testing across a range of groups.

8.3 The Four Themes of AI Safety

1. Alignment: the goals of the AI = the goals of people.
2. Robustness: resistance to malfunction and attack.
3. Interpretability: understanding the reason for a decision.
4. Control: a human can stop it.

2024~2025: Anthropic, OpenAI, and Google DeepMind all run safety teams.

8.4 Designing for Privacy

• Differential Privacy: add noise to the aggregate.
• Federated Learning: train without gathering the data.
• Homomorphic Encryption: compute while it stays encrypted.
• On-Device: the Apple approach.

8.5 Dark Patterns

• Confirm-Shaming: "Are you sure you do not want to subscribe? 😢"
• Forced Continuity: making cancellation difficult.
• Roach Motel: easy to sign up, hard to leave.
• Misdirection: pulling the eye somewhere else.

An engineer has an ethical responsibility to refuse to help design a dark pattern.

9. The Value System an Engineer Should Develop

9.1 Choosing Your Core Values

Answer for yourself: "what will I never trade away?"

• Honesty.
• Craft.
• User wellbeing.
• Transparency.
• Fairness.
• Autonomy.
• Growth.

Pick three to five, and write them down.

9.2 The Moment Values Conflict

• Company goals vs. user benefit.
• Shipping fast vs. quality.
• Personal promotion vs. team success.
• Short-term revenue vs. long-term health.

The fix: decide your own value hierarchy in advance. Refer to it when the conflict comes.

9.3 Saying No

• No to "this feature harms users."
• No to "this data collection is excessive."
• No to "this pace means compromising quality."

No has a cost (political capital, relationships). But a yes without values is a lifetime of regret.

9.4 Whistleblowing

• The extreme case: when the organization is causing serious harm.
• The AI ethicists who left Google before Duplex shipped.
• Frances Haugen at Facebook.
• In Korea: the protections are still weak and the personal risk is large.

10. Recommended Books by Topic

10.1 Introductory

1. "Sophie's World" — Jostein Gaarder: the history of philosophy as a novel.
2. "Justice" — Michael Sandel: the Harvard lectures on ethics.
3. "The Consolations of Philosophy" — Alain de Botton.

10.2 Epistemology

4. "The Logic of Scientific Discovery" — Popper.
5. "Thinking, Fast and Slow" — Kahneman.
6. "The Black Swan" — Taleb.

10.3 Ethics

7. "The Right Thing to Do" — Rachels.
8. "Practical Ethics" — Peter Singer.
9. "After Virtue" — MacIntyre.

10.4 Eastern

10. "Siddhartha" — Hermann Hesse.
11. "The Art of Living" — Epictetus / Sharon Lebell.
12. "Tao Te Ching" — Laozi.

10.5 Tech Ethics

13. "Weapons of Math Destruction" — Cathy O'Neil.
14. "The Age of Surveillance Capitalism" — Zuboff.
15. "Human Compatible" — Stuart Russell.
16. "Atlas of AI" — Kate Crawford.

11. Daily, Weekly, and Yearly Philosophical Training

11.1 Daily (10 minutes)

• Five minutes in the morning: Stoic meditation ("picture today's difficulties in advance").
• Five minutes in the evening: examination of conscience (did today line up with your values).

11.2 Weekly

• One philosophy essay or podcast a week (30 minutes).
• One ethical review of a technical decision a week.

11.3 Yearly

• Read one core philosophy book cover to cover each year.
• Revisit your values once a year.
• Refresh your tech ethics knowledge once a year.

12. The 12-Point Philosophy Checklist

• Three to five core values written down.
• The three ethical frameworks understood.
• Popper on falsifiability applied day to day.
• Bayesian updating as a habit.
• Stoicism practiced daily.
• Buddhist impermanence internalized.
• Dark patterns recognized and refused.
• Algorithmic bias checks as a routine.
• A quarterly values review.
• The muscle of saying no.
• One philosophy book a month.
• Following the latest trends in tech ethics.

13. The 10 Philosophy Anti-patterns

1. "Philosophy is abstract": it is a tool you use in everyday decisions.
2. "I trust my intuition": judging without verification.
3. "The company told me to": dodging moral responsibility.
4. "Technology is neutral": every technology has values built in.
5. "The data speaks": data too is biased by interpretation.
6. "If it is utilitarian it is fine": never checking whether the three frameworks converge.
7. "Zen means doing nothing": a passive misreading.
8. "There is only one right answer": monist dogmatism.
9. "Obeying the law is enough": legal is not the same as ethical.
10. "I am just an engineer": no sense of being a craftsman or a professional.

14. Closing — Philosophy Is an Ancient OS

"Philosophy is the art of living wisely." — Seneca

Programmers try to solve every problem with code. But the questions of why, right, and true do not yield to code. The ancient OS built for those questions is exactly what philosophy is.

For 2,500 years humanity has been debugging this OS. From Socrates, the Buddha, and Confucius through to Popper, Arendt, and Sandel. Only when this OS is installed underneath the tools an engineer uses every day does technology get pointed in a good direction.

In 2026, AI takes on more and more of the "how." What is left over is the "why." Why should this be built, for whom, and by which values.

Only the engineer who has been through those questions moves past Craftsman to become a Sage.

Three things are enough to start:

1. This week, write down three of your own core values.
2. In the next feature review, examine it through the three ethical frameworks.
3. Read one philosophy book this month (Sandel's "Justice" is a good pick).

Code runs without philosophy. But a life does not.

Next Post — "The Art of Relationships for Engineers: Family, Spouse, Friends, Children, Parents, and Community"

If philosophy was the values, relationships are the practice of those values. The next post covers:

• The engineering of a marriage — Gottman and the Four Horsemen
• Raising children and attachment theory (Bowlby)
• Caring for aging parents — time, money, emotion
• The maintenance of friendships
• Community, religion, and civil society
• The isolation of male engineers, and what to do about it
• Generational conflict in immigrant families
• Human relationships in the AI era

The last thing you stack on top of money, technology, and career. The final axis of the series, continued in the next post.