Why Logic Matters

Цели урока

• Understand why knowledge without logic is useless
• Learn what logic is and how it differs from common sense
• Learn to distinguish logic from rhetoric
• See where logic is needed in everyday life

1994. Intel ships the Pentium processor. One boolean error in the microcode's floating-point division logic. Recall cost: $475 million. 2017. An Amazon engineer types a console command slightly faster than intended. A single mistyped argument removes too many S3 servers. AWS goes down for four hours - taking half the internet with it. Therac-25, a radiation therapy machine, killed patients because of a race condition in a boolean condition. Logic is not an abstraction from a textbook. It is an engineering discipline with a price measured in human lives and nine-figure losses.

• **Formal verification** - the logic of safety-critical code is verified mathematically in aviation (ARINC 653), nuclear systems, and Tesla autopilot. One logical bug in a safety-critical system equals a catastrophe
• **LLM hallucinations** are logical errors in model reasoning. Chain-of-thought prompting (Wei et al. 2022) works precisely because it forces the model to build explicit logical steps instead of jumping to answers
• **SQL injection** remains #1 on the OWASP Top 10 not because developers don't know the syntax - but because they don't reason about the logic of trusting input data
• **A/B tests** - without understanding the logic of statistical inference, companies make wrong business decisions on "significant" p-values. Millions are lost this way every year

Paradox: knowing a lot != thinking correctly

**Paradox:** A person can know thousands of facts - and still make absurd decisions. Why?

**Key insight:** Knowledge is data. Logic is the algorithm that processes it. Without a good algorithm, even the best data produces garbage output.

A million Wikipedia articles are useless without the ability to tell fact from interpretation. Volume of information does not equal quality of inference. The bottleneck is never the data - it's the reasoning applied to it.

What logic is (and what it is NOT)

**Logic** is the science of correct reasoning. It studies not *what* is true, but *how* one statement follows from others.

**Logic is NOT:** ❌ Common sense (it often contradicts it!) ❌ Intuition (logic is precisely what checks intuition) ❌ Mathematics (though they are related) ❌ A tool for proving anything one wants **Logic IS:** ✅ Rules by which conclusions are drawn ✅ A tool for checking reasoning ✅ A language for analyzing arguments

The main task of logic is not to say *what* is true, but to check *whether* we are reasoning correctly. Establishing the truth of facts is the job of science. Checking the correctness of conclusions is the job of logic.

Logic vs Rhetoric: spot the manipulation

**Rhetoric** is the art of persuasion. **Logic** is the art of correct conclusions. These are different things.

**Trap:** The most persuasive arguments are often the weakest logically. The stronger the emotional effect, the more carefully the logic deserves scrutiny.

Advertising, politics, social media - rhetoric masquerades as logic everywhere. This is why technically brilliant people get caught by the crudest manipulations outside their domain: domain expertise does not transfer into logical immunity.

Key Takeaways

• **Knowledge is data, logic is the algorithm that processes it.** Without the algorithm, even the best data produces garbage output
• **Logic studies not what is true, but how one thing follows from another** - exactly what a type checker and a formal verifier do
• **Rhetoric persuades through emotion; logic checks through structure.** The most convincing arguments are often the weakest logically
• **Intel FDIV, Amazon S3, Therac-25** - three reminders that logical errors kill more than just arguments in a debate

Next

Now we'll learn to distinguish propositions - the basic building blocks of logic

• Propositions — next lesson

Связанные уроки

• ml-01