Infinity, Paradoxes, Gödel Incompleteness & the Mathematical Multiverse | Lex Fridman Podcast #488

Lex Fridman talks with set theorist Joel David Hamkins about the nature of infinity, beginning with Cantor's discovery that some infinities are larger than others and walking through Hilbert's Hotel, the diagonal argument, and Russell's paradox. They examine the rebuilding of mathematics on set theory and the ZFC axioms, then turn to Gödel's incompleteness theorems, the halting problem, and the deep distinction between truth and proof. Hamkins lays out the independence of the continuum hypothesis and argues for his controversial set-theoretic multiverse view, in which there is no single true mathematics but many alternative universes reachable via forcing. The conversation also covers surreal numbers, infinite chess, the philosophy of mathematical existence, the P vs NP problem, and Hamkins's skepticism about LLMs for mathematical reasoning.

• Cantor's profound achievement was proving the set of real numbers is uncountable, a strictly larger infinity than the natural numbers, via the diagonal argument, establishing that there is more than one size of infinity.
  00:09:31
• Cantor proved a far more general fact than the uncountability of the reals: for any set whatsoever, its power set is strictly larger, the abstract core behind Russell's paradox, the halting problem, and Gödel's theorem.
  00:50:22
• Gödel's incompleteness theorems decisively refute Hilbert's program: no computably axiomatizable consistent theory containing arithmetic can answer all questions, and no such theory can prove its own consistency.
  01:16:07
• Hamkins presents the halting problem's undecidability via a diagonal argument and shows Gödel's theorem follows immediately, calling it the simplest proof of Gödel's theorem, requiring no Gödel sentence.
  01:36:31
• Hamkins argues for the set-theoretic multiverse: decades of forcing results are evidence there is no unique set-theoretic reality, and the continuum hypothesis can be switched true or false like a light switch between closely related universes.
  02:35:06
• Hamkins and a collaborator proved the halting problem has a 'black hole', a computable procedure decides almost every instance correctly, with the proportion approaching 100% as the number of states grows.
  03:03:30
• Hamkins co-authored work building infinite chess positions with ever-higher ordinal game values (omega, omega-squared, omega to the 4th); it is now known every countable ordinal arises as a game value.
  03:35:00
• Hamkins warns LLMs are designed to produce arguments that look like proofs rather than arguments that are proofs, making them a dangerous source of error in mathematical reasoning.
  03:40:31

• Galileo anticipated Cantor centuries earlier, observing the perfect squares can be put in one-to-one correspondence with all natural numbers, but threw up his hands at the apparent paradox.
  00:04:18
• Hilbert's Hotel illustrates that a completely full hotel with infinitely many rooms can still accommodate a new guest, an infinite bus, and even an infinite train of infinite cars.
  00:10:01
• The rational numbers, despite being densely ordered (a fraction always lies between any two), are still only countably infinite, the same size as the integers.
  00:21:50
• The most famous transcendental numbers include pi and Euler's number e, and Cantor proved that most real numbers are transcendental.
  00:25:31
• Russell discovered the contradiction in Frege's logicist system just as Frege's monumental work was at the printers; Frege gracefully acknowledged it in an appendix.
  00:59:36
• Hilbert was so committed to set theory he declared 'No one shall cast us from the paradise that Cantor has created for us.'
  01:02:47
• Tarski's disquotational theory of truth, 'Snow is white' is true if and only if snow is white, gave the first rigorous definition of truth in a mathematical structure, distinct from proof.
  01:21:23
• The continuum hypothesis was Cantor's lifelong obsession (posed late 1800s), made #1 on Hilbert's 1900 list, and only resolved as independent via Gödel (1938) and Cohen's forcing (1963).
  02:24:32
• John Conway's surreal numbers are generated from nothing by a single rule, Hamkins calls the birth of zero the 'Big Bang of numbers', and unify the reals, ordinals, and infinitesimals.
  02:49:59
• In Conway's Game of Life, whether a given cell ever becomes alive is computably undecidable, equivalent to the halting problem.
  02:58:12