There's never been a better time to go into math (by: 3blue1brown) [video]

Original Article ↗ Hacker News Discussion ↗

Claims that “there’s never been a better time to go into math” collide here with anxiety over rapid advances in AI that can already solve or help solve difficult mathematical problems. Commenters weigh whether tools like large language models will augment human mathematicians—shifting emphasis toward interpretation, pedagogy, and problem selection—or instead hollow out the craft, careers, and sense of meaning that come from doing the hard work personally. Underneath is a broader argument about automation of cognitive labor: whether AI will simply change what humans do, or eventually leave very few genuinely intellectual roles for people at all.

Reception of the video and math communication

  • Many praise the creator’s visual explanations and breadth of topics; others say the style feels like “math as spectacle” that doesn’t translate into exam or work performance.
  • Some educators report that students who use the linear algebra series do better and gain intuition; skeptics argue that motivated students would do better anyway.
  • Several comments stress the value of human enthusiasm and storytelling in teaching, which AI videos currently lack.

Educational value vs AI-generated content

  • Some say AI-made “slop” videos actually help them more by being blunt and goal‑oriented rather than carefully foreshadowed.
  • Others argue that if AI ever produced videos as insightful and engaging as top human creators, that would effectively imply AGI.
  • LLMs as tutors are seen as powerful, but motivation and curiosity remain unsolved; simply providing a “supergenius” explainer doesn’t make learners engage.

AI’s impact on math as a profession

  • Strong divide:
    • Optimists: AI is just another tool; it will expand the frontier, create new math, and increase demand for humans to understand, unify, and teach it.
    • Pessimists: automation of theorem proving and problem solving “kills the craft,” ending an “age of heroes” and turning mathematicians into machine minders.
  • Some claim this is especially harsh because many entered math for the joy of deep problem solving, not for downstream roles like explanation, curation, or “babysitting a machine.”

Analogy to software and other fields

  • Numerous parallels drawn to software development: coding as a craft displaced by agents that “annihilate” routine coding, yet increasing the value of systems thinking and social skills.
  • Counterpoint: for some (e.g., autistic developers), rising social demands make the field less viable.
  • Historical analogies: agriculture mechanization vs horses; disagreement over whether AI is just another mechanization or fundamentally different because it can design itself and outcompete humans at cognition.

Meaning, motivation, and identity

  • Many emphasize that for mathematicians and programmers, the process is core to meaning, not just the outcome.
  • There is explicit grief over losing a life’s work or passion to opaque machine solutions, especially when those solutions are driven by corporate competition rather than curiosity.
  • Others respond that attachment to manual methods is ego or nostalgia and that advancing knowledge should matter more than who (or what) gets the proof.

Economics, jobs, and society

  • Debate over whether AI expands the “pie” of work or just increases the number of “pie eaters.”
  • Fears that if AI can surpass humans across cognitive tasks, very few human roles will remain, leading to mass unemployment regardless of whether it wipes out 90% or 100% of jobs.
  • Some argue that past precedents (mechanized agriculture, prior automation) are not good guides, because humans could always “move up the stack” before; that may no longer be possible.

Governance and open questions

  • A minority proposes strong regulation: banning monetization of models trained on public data, enforcing royalties for style/IP copying, invalidating AI-generated IP.
  • Others push back, arguing users should be free to adopt AI tools; banning them would “freeze” society in the 2010s.
  • Open questions in the thread:
    • Will AI cap below human-level in some domains, or exceed us everywhere?
    • Can AI also do the “biology-style” work of explaining its own results to humans?
    • Is “going into math” today reasonable as a decades-long career bet, or mainly as a personal passion?