Ansh Krishna
Computer Science & Mathematics at Columbia University
ansh [dot] krishna [at] columbia [dot] edu

i spend a lot of time working on compute markets and the financialization of gpu hours. at columbia, i'm doing research in industrial engineering and operations research across two threads. one builds hierarchical reinforcement learning frameworks that override and optimize autonomous trading agents' positions in real time. the other recovers the pricing kernel from the divergence between prediction markets and crypto options written on the same events.

  1. mistakes of ambition age better than mistakes of sloth.
  2. speed is the truest test of intelligence.
  3. inaction is just missed shots.
  4. every acceleration becomes a constant, so you have to recalibrate often.
  5. the efficient market hypothesis is a lie.
  6. be kind, be direct.
  7. physics is the only real limit.
  8. the truest art is made at the research frontier.
  1. if the marginal cost of production asymptotically approaches zero across goods and services, driven by AI, robotics, and energy abundance, what happens to price signals, capital allocation, and the incentive structures that underpin market economies? does profit motive collapse, or does scarcity simply migrate to attention, status, and the irreducibly human?
  2. what are the upper bounds of recursive AI self-improvement? if a system can meaningfully accelerate its own research throughput, does capability growth become non-linear in ways we can anticipate, or does it hit hard walls in compute, data, or the structure of knowledge itself?
  3. if AI offloads enough cognitive and physical labor, what does the human body actually become? do we atrophy, cognitively, physically, socially, or does biology find new gradients to optimize against? what happens to embodiment when effort is optional?
pieces of literature worth carrying
  1. meditations — marcus aurelius
  2. bhagavad gita — vyasa
  3. the tao te ching — laozi
  4. the prince — niccolò machiavelli
  5. walden — henry david thoreau
  6. the symposium — plato
  7. confessions — augustine
  8. the iliad — homer
  9. the odyssey — homer
  10. inferno — dante alighieri
  11. invisible man — ralph ellison
  12. frankenstein — mary shelley
  13. jane eyre — charlotte brontë
  14. ficciones — jorge luis borges
  15. the beginning of infinity — david deutsch
  16. a brief history of time — stephen hawking
  17. the order of time — carlo rovelli
  18. the selfish gene — richard dawkins
  19. gödel, escher, bach — douglas hofstadter
  20. on intelligence — jeff hawkins
  21. thinking, fast and slow — daniel kahneman
  22. sapiens — yuval noah harari
  23. 21 lessons for the 21st century — yuval noah harari
  24. the lessons of history — will & ariel durant
  25. the rational optimist — matt ridley
  26. the black swan — nassim nicholas taleb
  27. antifragile — nassim nicholas taleb
  28. the bed of procrustes — nassim nicholas taleb
  29. poor charlie's almanack — charlie munger
  30. the intelligent investor — benjamin graham
  31. innovator's dilemma — clayton christensen
  32. the innovation stack — jim mckelvey
  33. zero to one — peter thiel
  34. the hard thing about hard things — ben horowitz
  35. shoe dog — phil knight
  36. what it takes — steve schwarzman
  37. the money trap — alok sama
  38. mind without fear — rajat gupta
  39. an era of darkness — shashi tharoor
  40. surely you're joking, mr. feynman — richard feynman
  41. the lean startup — eric ries
  42. the alchemists — neil irwin
  43. diamonds in the dust — saurabh mukherjea
  44. an economist goes to the game — paul oyer
  45. relentless — tim grover
  46. atomic habits — james clear
  47. ikigai — hector garcia & francesc miralles
  48. lying — sam harris
  49. this is water — david foster wallace
  50. how to think for yourself — paul graham
  51. is it worth being wise? — paul graham
  52. do things that don't scale — paul graham
  53. situational awareness — leopold aschenbrenner
  1. twitter
  2. github
  3. linkedin
  4. instagram