The Problem of Teaching Physics in Latin America (1963)

Feynman’s diagnosis and its generalization

• Commenters see Feynman’s Brazil experience as a special case of a universal issue: students learning to recite definitions and pass exams, not to understand or apply concepts.
• The focus on credentials and “productive workers” is contrasted with genuine learning; credentials are seen as gatekeepers to jobs rather than markers of competence.

Rote learning, credentials, and assessment

• Many recall exams that rewarded recall rather than reasoning, and only “learned physics” when building or breaking real things.
• Others report the opposite: open‑book, problem‑solving exams where most students still failed, suggesting assessment design strongly shapes what students optimize for.
• Goodhart’s law is invoked: once grades and diplomas become the target, systems optimize for test performance, not understanding.

AI/LLMs and the same old problem

• Some argue LLMs worsen Feynman’s problem: teachers can auto‑generate content they don’t understand; students can auto‑generate homework, further divorcing credentials from knowledge.
• Others say banning AI is unrealistic; better to treat outputs as hypotheses or drafts and design exams (oral, in‑person, problem‑solving) that require independent thinking.
• There is disagreement on whether AI will “wreck” the current education system in a good or bad way.

Teaching for understanding

• Suggested practices: non‑copyable exam questions, reduced curriculum breadth in favor of depth, frequent problem‑solving in class, and emphasizing intuition and geometric/conceptual models over symbol‑pushing.
• Several educators stress that students must ultimately “do the snowboarding” themselves, but institutions can strongly incentivize understanding instead of memorization.

Mass education, inequality, and institutions

• One line of argument: as education scales to the whole population, quality and teacher expertise inevitably drop; elite models don’t transfer directly to mass systems.
• Class size, funding, corruption, and rigid bureaucracy (e.g., difficult course transfers) are cited as structural barriers.
• Another thread asks how to sort students into appropriate levels and allow mobility as their performance or interests change.

Latin America, economics, and geopolitics

• Some see low salaries and weak science institutions as simple consequences of poverty; others argue there is “enough money” but misallocation and corruption.
• The “international division of labour” is blamed for trapping some countries in primary-goods extraction while manufacturing nations capture most of the gains and improve education.
• A heated subthread debates whether US/Western intervention and coups are central to Latin America’s underdevelopment, versus internal responsibility and local governance.
• There is also pushback that Feynman’s 1960s snapshot no longer fits all countries; examples are given of modern Latin American systems (e.g., Uruguay) with strong problem‑solving cultures and global‑level graduates.

Attitudes toward physics and career incentives

• In some regions, physics is high‑prestige and chosen for love of the subject; in others it is what you study if you “couldn’t get” engineering in rank‑based systems.
• Rank and prestige can push bright students away from their interests (e.g., physics) into more lucrative or status‑heavy fields, potentially harming both learning and long‑term fulfillment.

Everyday intuition and real‑world physics

• Multiple anecdotes highlight the gap between knowing formulas and seeing mechanisms in daily life (e.g., hot water lag in pipes, component tolerances in circuits).
• These are used to illustrate Feynman’s key point: real understanding is the ability to connect abstract knowledge to concrete phenomena, not just to recite laws.