Why The Search Has To Be Expensive

Every passage, on the record.

1. 00:32Analogy

   You think about Albert Einstein, when he is trying to figure out E equals MC squared, it took him years to do that. But now, any physicist, any physics student can verify the same exact math in an hour.

2. 01:04Claim

   There's a particular asymmetry that runs through every active knowledge generation in the universe. Finding the right configuration, finding the right arrangement, is harder than checking it to see that it's right. That means the search has cost. We are searching phase space, looking for a solution, the right arrangement. Once we find the right arrangement, the verification is cheap.

3. 01:37Reference

   It has a name in mathematics, a name that pretty much anyone outside of computer science has never heard of. That's P versus NP. The claim of this episode is that P versus NP isn't just a technical question about the limits of computers, it's the cost asymmetry that runs through every active knowledge generation at every scale in the universe.

4. 02:13Claim

   The universe has been running this engine for its entire existence. It's in the stars, it's in the genomes, it's in the minds, it's in the scientific community, it's in published books, it's in functioning civilizations, it's in memory. For the past 16 years it has been running publicly, openly on the internet, inside the computer network we call Bitcoin. Where the asymmetry holds, knowledge accumulates and time is generated.

5. 02:43Idea

   Information is raw possibility, constraint gives possibility a shape. Knowledge is constrained information with reach. And time is the rate at which knowledge generation opens previously inaccessible regions of phase space. And it does that by paying in heat.

6. 03:17Reference

   Last episode made the case about what time is, physically. The problem, or what was missing, was a name for the cost asymmetry that the engine itself runs on.

7. 04:22Analogy · condensed

   Take the super easy example of a puzzle. A thousand piece puzzle is going to take you a long time, you have to check which pieces fit where. For verifying whether or not it's done, it's almost instantaneous, one second, bang, it's super cheap.

8. 04:56Analogy

   Solving was hard and checking was easy, takes you days and days to assemble and seconds to verify. Same thing going on with the Bitcoin block: you get on average ten minutes of enormous industrial computation to mine a block, and it takes milliseconds to verify on your little Raspberry Pi.

9. 04:56Quote

   It takes milliseconds to verify for your little Raspberry Pi that you've got in your closet, that is defending you against the Fiat overlords.

10. 06:00Claim

    NP stands for non-deterministic polynomial time. The important property is: if somebody hands you a candidate solution, you can check it super quick.

11. 06:31Claim

    The other side of this asymmetry is the family of problems where finding a solution is itself quick, like sorting a list of names alphabetically, multiplying large numbers, or searching a database for a particular entry. These are problems where solving and checking take comparable effort. Mathematicians call this family P, for polynomial time.

12. 07:06Claim

    Almost every working mathematician or computer scientist believes finding is genuinely harder. They believe P does not equal NP. Nobody has proved it. The smartest people in the world have been trying for 55 years, but this is not a technicality.

13. 07:43Claim

    If you walked in tomorrow with a proof that P equals NP, the world's digital infrastructure could collapse overnight. Public key cryptography, signatures on your banks, all the HTTPS internet connections, your messaging app, your password manager, they all become easily breakable. The same could be said for Bitcoin: mining would cease to be a search problem, and blocks would generate as fast as anyone could request them.

14. 08:20Claim

    The smart money is on P not equaling NP, and on the digital security infrastructure of civilization resting, at least for what we know now, on sound mathematical ground. But the proof hasn't arrived, and until it does, the asymmetry remains a conjecture. It's a very well supported conjecture, but a conjecture all the same.

15. 08:54Claim

    The same asymmetry appears in every system that generates knowledge, not just cryptographic protocols. Once you see this, you can't unsee it. The asymmetry is everywhere, and it has the same shape everywhere. The shape has consequences.

16. 09:28Idea

    His brain is a container with finite contents and an unbounded interaction surface. Inside the finite contents are however many billions of neurons, configured by 26 years of life, by formal training in physics and mathematics, by the embedded experience of working as a patent clerk reading patents and clock synchronization. The unbounded interaction surface is everything he is still pressing against.

17. 10:13Idea

    Einstein's brain has high constraint quality. The interior is rich with established principles that have been tested against reality his entire life. Maxwell's equations have been verified against every experiment since the 1860s. Newtonian mechanics has been verified against every projectile, every planet, every pendulum, for two centuries.

18. 10:48Claim

    These are not soft beliefs. They're durable structural patterns inside his neurons with massive verification behind them. Einstein is searching the space of possible mathematical relationships for a configuration that satisfies both Maxwell and Newton.

19. 11:20Claim

    The search is exactly the kind that NP describes. Finding the answer is hard, very hard. Once found, the answer gets really, really quick to verify.

20. 11:53Idea

    He knows the answer has to be a Lorentz invariant. He knows it has to reduce to Newtonian mechanics at low velocities. He knows it has to conserve energy. He knows it has to make experimentally testable predictions. These constraints turn an astronomical search space into a tractable one.

21. 12:29Claim

    That is what C squared actually does in K equals I C squared. It's not just verification quality, it is search pruning power. The richer your interior of established constraints, the more efficiently you can search inside an NP asymmetric space.

22. 13:03Claim

    Most of today's searches in physics are too hard to do without much richer interiors than any single human can hold. That's why we get theoretical physics happening in groups or over decades now. The interior has to grow before the next search becomes feasible. This is what scientific institutions are ultimately for.

23. 13:03Idea

    The moment Einstein finds the equation, the pawl catches in his brain. The pawl is the little part of the ratchet that doesn't allow it to go to release. It now stays in that configuration.

24. 13:33Claim

    He pays the energy of the search, the months of metabolic effort burned in his neocortex. The neural pattern that encodes E equals MC squared is true gets laid down irreversibly. The verdict comes from external reference, the constraints he already trusts: he runs the equation through Maxwell, through Newton, and through the thought experiments he's been refining for years.

25. 14:05Quote

    It passes all of those things. He can't unthink it. There's no path back to a state of pre-knowledge.

26. 14:05Claim

    That is one tick of internal time generated in one brain, one block in the local chain. But now the horn branches: Einstein doesn't keep this knowledge to himself, he writes the paper, and the neural pattern in his brain gets copied to a new substrate, ink on paper. The paper is a new horn.

27. 14:37Idea

    Its interior is finite, a few pages of mathematics and some prose, but its boundary is potentially unbounded. Every reader, every translator, every textbook author who will ever interact with it.

28. 15:07Claim

    The act of copying the pattern from brain to paper is energetically expensive. He has to translate from the way the idea lives in his neurons to the way it lives in symbols, and the translation is lossy in some directions and crystallizing in others. The symbolic form is much more durable and transmissible than the neural form, but it loses some of the embodied intuition Einstein has that the reader doesn't. The horn that branches off is a compressed version, optimized for transmission.

29. 15:39Quote

    The brain horn can't move. So it makes a paper horn that can.

30. 16:10Claim

    Planck's brain horn presses on the paper horn. He verifies the math, he checks the consistency with what he already knows. The verification is getting cheaper and cheaper: hours or days versus the months and years it took Einstein.

31. 16:41Claim

    The asymmetry between the years Einstein spent on the equation and the hours Planck spends checking it is the cost asymmetry of NP at work, just made visible at a human scale.

32. 17:11Claim

    Each check is a tick of that pawl catching in a new brain. The horn's constraint is climbing faster and faster and faster. And that's when the experiments start.

33. 17:46Claim

    The horn is so well constrained that nobody seriously doubts it. It's no longer being tested at a basic level, it's being used as a constraint that prunes the search space for the next generation of theories.

34. 17:46Quote

    A brain horn pays an enormous search cost to find a configuration. The configuration gets copied to a paper horn at the cost of some compression. The paper horn enters a community of other horns.

35. 18:26Analogy

    Darwin spent decades searching the space of possible explanations for the diversity of life. His C squared was high because of his geological training, his reading, his correspondence with his friends, his five year voyage. His prior constraints pruned the search. He found natural selection, one specific point in phase space, inaccessible before he found it, accessible after. The notebook he wrote in was the horn. The Origin of Species was the bigger horn. The community of naturalists who read it, bigger horn again.

36. 18:26Reference

    Darwin spent decades searching the space of possible explanations for the diversity of life.

37. 18:26Reference

    The origin of species, the bigger horn, the community of naturalists who read it, again bigger horn.

38. 19:05Claim

    When you trace the horns through the levels you see this fractal structure of knowledge generation. A neuron firing pattern is a horn. The thought that the firing pattern instantiates, the sentence that expresses the thought, the page that records the sentence, and the book that contains the page. At every level, the same asymmetry holds.

39. 19:37Claim

    Inside any container that has time, search is more expensive than verification. This is what makes the container generate time in the first place. If P equaled NP at any level, if finding became as cheap as checking, that level would collapse into a static dream. Knowledge would not accumulate there because there would be no cost asymmetry to generate time.

40. 20:12Quote

    The reason civilization has gotten richer in knowledge across centuries is that P does not equal NP.

41. 20:48Analogy

    A giant miner is drawing a certain number of megawatts of electricity from the grid, with a warehouse of hundreds of thousands of specialized computers all doing one thing, guessing numbers. They'll guess for, on average, about ten minutes, and make a metric fuck-ton of guesses before one of them gets lucky and finds a number with the right property.

42. 21:20Claim

    Within seconds, every full node on the network checks the lucky guess. The check is a single hash, milliseconds. But a miner had to burn an industrial city block worth of electricity to find that number. Every node confirms it in the time it takes to say this line.

43. 22:26Quote

    Bitcoin is the cleanest engineered instance of this universal pattern. It's the smallest, it's the simplest, it's the most observable case of an NP asymmetric container generating its own internal time.

44. 23:00Quote

    Money is the first application of the machine. It's the simplest one, it's the one with the clearest market bid. But the underlying machine is a time generator.

45. 23:00Claim

    Each block is one quantum of inaccessible phase space being made accessible. Each block is paid for in real watts, dissipated as real heat. The chain that accumulates is a record of every tick of internal time that the protocol has generated since inception.

46. 23:34Idea

    The shape of Gabriel's horn is the shape of an NP problem, that's what it looks like. The interior is the verifier, the boundary is the search space.

47. 24:06Quote

    The horn is not a metaphor for NP. The horn is the geometric picture of the cost asymmetry that NP names in mathematics.

48. 24:44Analogy

    The bucket of paint that fills the interior is the rules of the system. The infinite surface is everything you have to search through to find a valid solution. When Einstein is searching for that equation, he's traversing the surface of the horn. When Planck is verifying it, he's measuring the volume of the horn. The discovery is the painting of the boundary. The verification is the filling of the interior.

49. 25:18Claim

    Mining is a search problem with a verifier. The miner is trying to find a number such that, hashed together with the rest of the block, the result begins with a long run of zeros. The space of possible nonces is enormous, but only a small fraction of nonces produce a valid block.

50. 25:49Claim

    There's no clever shortcut, no algorithm that lets you skip the guessing. That asymmetry is the security. The cost of generating a valid block is many millions of times higher than the cost of verifying one. Bitcoin security rests on the belief that P does not equal NP, and that the cryptographic hash function Bitcoin uses has no shortcut.

51. 26:20Claim

    If P were proven to equal NP, the whole search and verify architecture of Bitcoin loses its mathematical floor and mining becomes very cheap. The chain would still exist, but it wouldn't be secured by the cost of mining. The asymmetry vanishes.

52. 26:20Claim

    Even if the computational asymmetry vanished, Bitcoin itself would not be fully refuted. It would lose one of its two pawls, the two things that make sure it doesn't go backwards, and the other one would survive.

53. 27:22Idea

    There's actually two pawls, and Bitcoin sits on top of both of them. The first is thermodynamic, that's Landauer's principle: every bit of information you store, every bit you erase, every bit you transform irreversibly has an energy cost that can't be reduced below this floor. That pawl doesn't depend on whether P equals NP, it doesn't depend on cryptography, it's a property of the universe. The second pawl is the computational part, and that one's the conjecture, the asymmetry between P versus NP.

54. 27:22Reference

    The first is thermodynamic, right? That's Landauer's principle.

55. 27:52Quote

    Information is physical. Erasing information does cost energy.

56. 28:22Claim

    The energy dissipation is forced by Landauer, the hash difficulty is forced by the cryptographic conjecture. That's the two pawls bolted together. It's the only public machine humans have built that bridges both pawls at scale.

57. 29:23Claim

    Each tick of internal time is one quantum of inaccessible becoming accessible, paid for by an irreversible state change. The conversion only has any cost, and therefore only generates any time, if the search side is genuinely harder than the verification side.

58. 29:55Claim

    If finding configurations were as cheap as checking them, the inside of the container would already be accessible to itself, every configuration would already be present. Time in this framework exists in the regime where P does not equal NP holds for the interior of a container.

59. 30:27Claim

    A container with no cost asymmetry inside has no time. From outside the container, looking in with an oracle that grants instant access to any configuration, the cost asymmetry doesn't exist. P equals NP from that view, trivially, because there is no search, no cost to anything, no time either.

60. 30:57Claim

    From inside the container with no oracle access, configurations have to be searched for. The search has a cost. That cost is what we have always called effort or work or energy expenditure or irreversible commitment. The accumulation of those costs across the container's history is what we have always called time.

61. 31:33Idea

    Wet clay is information. It can take any shape. You can mash it, you can reshape it, you can copy it, you can smash it. Nothing sticks, nothing has to be paid for. There's no constraint, there's no kiln, there's no pawl, there's no irreversibility. That's what cyberspace is right now. Bits are just clay.

62. 32:05Idea

    Cyberspace is a giant clay pile that anyone with admin rights can then reshape. Bitcoin is a kiln inside of cyberspace. The pawl is the proof of work mechanism. The mining process is the firing. To produce a Bitcoin block, a miner has to pay real energy, dissipate real heat and run through that four step mechanism from episode 187.

63. 32:05Reference

    A miner has to pay real energy, dissipate real heat and run through that four step mechanism from episode 187.

64. 32:43Claim

    You gotta commit energy upfront, make a transformation that is irreversible, and then the verdict comes from an external reference the miner doesn't control, the network of nodes. There's no path back. That is the kiln, running inside a domain that previously had no kilns.

65. 33:14Claim

    The miner has to do trillions of guesses, each one dissipating heat. That hardness is P does not equal NP. If P equals NP, the kiln is gone, it cools down, finding the right hash becomes super cheap, the four step mechanism stops altogether. The pawl doesn't catch, the clay never bakes.

66. 33:48Idea

    What's the opposite of a horn? Well, the opposite of a horn is a cloud. Cyberspace is full of clouds. It's the inverse topology of a horn. It's a bounded surface, which is like the company servers, wrapped around an unbounded interior, which is whatever data they want to store, edited or changed at will.

67. 34:22Claim

    Nobody is pressing on the boundary of a cloud, nobody is verifying it. The cloud's owner can rewrite the interior at any time and you'd never know. A cloud has no real interior, it has a database somebody else owns. It can't generate constraint quality because there's no external verification grinding against it, and there's no horn ecology because there's no exhaust being passed between independent horns. It's just a static blob.

68. 34:53Claim

    Bitcoin is the first horn in cyberspace: it's got the finite interior, it's got the unbounded boundary, and the exhaust from each block becomes the fuel of the next block. It is the horn ecology in digital form. P does not equal NP is what lets the horn be a real horn instead of collapsing into a cloud.

69. 35:24Quote

    Cyberspace stays clay, stays cloud with no record. But if you put it in there, well then it becomes baked, becomes a horn.

70. 35:59Reference

    This goes back to Szabo and Jason Lowery. You know how humans settle disputes in physical space: they project power, they show up with watts, whether the dispute is a war or a court case or a property line or an estate battled between siblings. It eventually settles when one party makes a commitment the other can't undo.

71. 35:59Reference · condensed

    This goes back to Szabo and Jason Lowery.

72. 36:29Claim

    That's the pawl catching, that's the commitment sticking. The dispute is resolved by an irreversible state change in physical reality. That mechanism doesn't work in current cyberspace: anything can be edited, faked, copied, redacted, so disputes in cyberspace can't really be settled, they just get deferred until somebody takes them back to physical reality where the commitment actually sticks. Bitcoin introduces that irreversible state change.

73. 37:03Claim

    P does not equal NP is the thing that makes the cost real. Without it, the cost of mining collapses and that irreversibility disappears. Virtual reality only becomes a place where things matter if it has the same dispute settlement architecture as physical reality.

74. 37:03Reference

    This also works with what we talked about in episode 191 of the beginning of time, being the state with zero accessible phase space, because that's what the beginning of time actually means.

75. 37:39Claim

    The beginning of time is the state where the container is so small, so constrained, that the asymmetry between inaccessible and accessible hasn't yet opened. Everything inside is fully accessible to itself because there's nothing inside to be accessed. P equals NP holds at time zero, trivially, because the search base is empty.

76. 38:14Claim

    The start and end of time correspond to the two trivial regimes where the asymmetry does not hold. Time is that middle region. This cost structure is what makes Bitcoin a time generating machine. Somebody like Jason would argue Bitcoin is a power projection machine. That's true, but Bitcoin is also a time generating machine, and the second framing shows something the first doesn't.

77. 38:46Quote

    Bitcoin participates in the same physical project the universe has been engaged in forever. It's forging durable constraint by paying irreversible costs to traverse asymmetric search spaces.

78. 39:17Claim

    It's realized in the physical substrate of the universe because of Landauer's principle. The horn doesn't have to be just a picture, it is the shape that knowledge has to take in any container where the cost asymmetry holds.

79. 39:52Idea

    The universe has a preference for horns. Not for some mystical reason, but because the joint constraint of information being physical and verification being asymmetric carves a specific path through configuration space. Any structure that finds that path persists, while any structure that does not dissipates. The horn is the attractor.

80. 40:22Claim

    This gives us a design principle for institutions, for protocols, for organizations. It gives us a way to evaluate whether a proposed structure is geometrically sound before it has time to fail. It also gives us a test on AI, whether it's generating new knowledge or just laundering existing knowledge.

81. 40:22Quote

    It also gives us a test on AI, what they're doing, whether it's generating new knowledge or just laundering existing knowledge.

82. 40:54Open question

    We're sort of reframing evolution itself as selection for horn geometry against a universal cost asymmetry. So that's what I'm gonna talk about in the next episode.