newsfilter.io
Interview, Fireside Chat

Po-Shen Loh: Mathematics, Math Olympiad, Combinatorics & Contact Tracing | Lex Fridman Podcast #183

  • Plans to release short educational videos covering AI, robotics, math, science, philosophy, and life snippets over the coming months, with potential expansion into music, books, and martial arts based on interest.
  • Anticipates society will require pandemic solutions similar to Novid in the future, with a hope that global responses will achieve better economic stability and support for entrepreneurs than the initial COVID reaction.
  • Intends to collaborate with public health researchers at various universities to conduct pilot deployments of the Novid system, which aims to serve as an end-game strategy for COVID once cases are low and allow societal reopening.
  • Expects Novid to be beneficial for regions unable to afford expensive vaccines like Moderna or Pfizer (priced at approximately $20 per dose) and predicts that future vaccine makers are preparing for perpetual new doses.
  • Predicts that Novid will maintain user anonymity while preventing manipulation through verified codes issued by local health authorities, scaling effectively due to sparse social networks where most individuals have at most 100 strong connections for linear-time computation.
  • Hopes that Novid's approach of incentivizing self-protection will outperform current contact tracing methods and create a viral feedback loop by showing users they are indirectly connected to hundreds or thousands of others.
  • Anticipates that insights gained from solving Novid and pandemic problems will outlast multiple generations, with mathematical techniques generalized to understand social networks, political bubbles, and recommender systems.
  • Expects that solving complex problems generates a "thinking muscle" and discipline, and that students experiencing the "power to invent" through difficult math may fall in love with mathematics; coaches hope students will have the energy to solve major catastrophes by 2030.
  • Believes that human intelligence will evolve to value knowledge over resources, and that space travel will eventually be as commonplace as air travel, while AI borrowing these ideas could lead to the creator being remembered as a civilization preserver.
  • Predicts that if mathematics is used to communicate with alien species, the shared curiosity will outweigh desires for resources or violence, and hopes methods for solving equations could reach trillions of people if humanity becomes multi-planetary.
  • Expects the "voting tree" circuit to improve from square root of n to a better approximation, while a modified stochastic coalescence algorithm accepting the smallest lump should achieve more efficient time complexity.
  • Suggests distributed algorithms may form representative government structures organically from the bottom up and notes P likely does not equal NP, with Scott Aaronson's 3% probability of P equaling NP motivating many researchers.
  • Hopes that a quadratic equation method discovered has a small chance of entering textbooks to increase long-term impact, while predicting humans will make their own risk decisions using apps like Novid even if vaccinated.
  • Expects that mathematical insights derived from pandemic control could have indirect impacts on understanding human interaction networks unrelated to disease and that a daily practice of thinking about non-work problems builds discipline.
  • Believes teaching a problem to children ensures the teacher learns it, and that mathematically solving a complex problem yielding a dead end still provides value through the process.
  • Hopes that the "voting tree" circuit he studied will be improved from square root of n to a better approximation, though not necessarily n/2, and that distributed algorithms can form representative government structures organically.
  • Believes that the stochastic coalescence algorithm, if slightly modified to accept the smallest lump, will achieve a more efficient time complexity than the original square root of n, and expects that the quadratic equation method he discovered has a small chance of making it into textbooks.