Lec 16 Pedagogy in the Times of AI
Capitalist abundance widens economic inequality
Citing arguments by Geoffrey Hinton, PR highlights how cheap, commoditized intelligence under capitalism will not automatically generate universal well-being. Instead of solving poverty, unchecked production will exacerbate wealth disparities between compute monopolies and the broader public because systemic human greed remains unresolved. PR concludes that compute concentration creates modern feudal lords who dictate societal resource distribution.
Collapse of meritocracy and professional labor
PR warns that the conventional economic model rewarding intellectual skills acquired through rigorous education faces collapse within five years. Privileged technical graduates face corporate downsizing as human researchers become organizational bottlenecks compared to parallel autonomous agents. PR concludes that educational institutions and policymakers are entirely unprepared to address widespread displacement across developing economies like India.
Industrial revolution analogies fail to apply
PR disputes student suggestions that current AI disruption mirrors the historical Industrial Revolution, noting that earlier transitions only mechanized physical labor while expanding the knowledge economy. Replacing cognitive labor leaves displaced humans without an obvious substitute economic domain. PR asserts that mere welfare support or endless digital dopamine cannot satisfy human psychology, making civil unrest plausible if meaningful employment evaporates.
Human motivations for physical classroom attendance
PR questions why students continue attending university lectures in person given that large language models offer infinite patience, clearer technical explanations, and superior customized tutoring. Reflecting on his own transition from watching online university lectures to querying models directly, he asks whether classrooms offer social connection or mere habit. He concludes that passive content delivery has lost all distinct utility.
Classrooms reimagined as cognitive workout gyms
Drawing on Indian philosophical traditions of one-on-one mentorship, PR proposes that modern university classrooms function less as lecture theatres and more as cognitive gymnasiums. Just as individuals join physical fitness clubs for external discipline and community motivation rather than unique equipment, students attend universities to maintain rigorous intellectual habits. He concludes that institutions must cultivate general analytical competence rather than memorizing ephemeral algorithms.
Interactive problem sessions replace rote syllabus
PR explores restructuring class time into equal halves, dedicating 45 minutes to self-directed AI research on specific technical challenges like diffusion model losses followed by 45 minutes of interactive faculty debate. However, he acknowledges that scaling authentic one-on-one mentorship across 100 students remains unsolved, even with student-developed AI personas trained on his past lectures.
Business shifts from tech to perceived value
PR analyzes how individual workers must transform into independent value creators in an automated economy, noting that commercial success depends on shaping the perception of value rather than engineering technical artifacts. He remarks that philosophy often trapped thinkers in endless contemplation without action, whereas engineering demands practical resolution. The discussion resolves that educators must guide students in identifying genuine human needs and convincing markets of their utility.
Rigid grading curves penalize authentic education
PR criticizes mandatory institutional grading policies that enforce Gaussian bell curves across motivated classes, sharing a story by Kannada satirist Beechi comparing compulsory schooling to dragging livestock to slaughter. He laments that forced attendance and mathematical grading metrics alienate students rather than fostering mutual trust.
