Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Magic Internet Math

Magic Internet Math artwork

Perfect Number Day and the Road to Schnorr

Jun 29, 20261 hr 10 minAudio

We celebrated June 28 as “Perfect Number Day” and finally told one of my favorite origin stories behind this whole project: a late-night drive home from a Phish show with Kayla, where a quick conversation about why 6 and 28 are perfect numbers turned into a full-blown family math obsession. We revisited how that curiosity led us to trial-and-error the next perfect number, fire up a Raspberry Pi, write Python code, hit the limits of the machine, and eventually discover just how deep number theory goes - from the Pythagoreans to Euclid, Mersenne primes, and the still-open questions around perfect numbers. From there, we closed out our elliptic curve study guide by turning to Schnorr signatures, Taproot, MuSig, FROST, libsecp256k1, and the practical realities of Bitcoin development. We talked through why Schnorr is cleaner than ECDSA, how signature aggregation improves privacy and scalability, why wrench-attack resistance matters, and why the ecosystem around libsecp256k1 deserves far more attention. We also reflected on the tiny, careful pace of high-stakes Bitcoin cryptography work, the importance of maintainers like Jonas Nick, Pieter Wuille, Tim Ruffing, and Andrew Poelstra, and why understanding these mathematical foundations is the whole point of Magic Internet Math. Magic Internet Math: https://magicinternetmath.com/ Magic Internet Math Podcast: https://podcasts.apple.com/us/podcast/magic-internet-math/id1868224151 Phish Tour Archives: https://phish.com/tour-archives/ BIP 340 - Schnorr Signatures for secp256k1: https://bips.dev/340/ BIP 341 - Taproot: SegWit version 1 spending rules: https://bips.dev/341/ BIP 342 - Validation of Taproot Scripts: https://bips.dev/342/ BIP 327 - MuSig2 for BIP340-compatible Multi-Signatures: https://bips.dev/327/ RFC 6979 - Deterministic Usage of DSA and ECDSA: https://www.rfc-editor.org/info/rfc6979/ RFC 9591 - FROST Threshold Schnorr Signatures: https://www.rfc-editor.org/info/rfc9591/ libsecp256k1: https://github.com/bitcoin-core/secp256k1 rust-bitcoin: https://rust-bitcoin.org/ AnchorWatch: https://www.anchorwatch.com/ Blockstream: https://blockstream.com/ Blockstream Research: https://research.blockstream.com/ Jonas Nick: https://blog.blockstream.com/author/jonas/ Andrew Poelstra and the Blockstream Research team: https://blog.blockstream.com/blockstream-research-the-focus/ Pieter Wuille (sipa): https://github.com/sipa Daniel Shanks - Solved and Unsolved Problems in Number Theory: https://bookstore.ams.org/chel-297/ SHAttered (the SHA-1 collision project): https://shattered.io/sha1-collision/

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

UnknownOkay. We're back. We are back at the Magic Internet Math podcast. I am your cohost, Brian Hirschfield, and I have Rob Hamilton. And we have a guest today named, you may know from the last episode. It is, Kayla. Kayla, say hello. Hi. Alright. We're all here. We're all here. We're all here. I'm introducing us now every time. Get it. The clankers have to get things right. Yeah. You have to set the context up.

Exactly. We're here, and it's a very special day today. And there's a reason we're rolling definitely today, and there's a reason Kayla's here. And the reason is today is, today is the true math day on the calendar to me. Like, pie day is the one that exposes all the larks of, like, oh, I love numbers and all that. Right? Like, that's the day you know, I thought it would be funny. I don't if I'm the only one on Earth who thought this was funny, but, like, I thought that,

what was it, that July 22 in Europe would be a good day for Pi Day because they do the day first and then the month. So it's twenty two sevenths. Twenty two slash seven.

UnknownOh, yeah. For the prime approximation. Yeah. Yeah. Yeah. Would be more of, an if you know, you know kind of pie day. Yeah. That was always the inside baseball pine pie approximator. 22 divided seven. 22 divided by seven. Yes. That's like yeah. And whenever I was doing a math test in high school, I can do that on my calculator. I'm old. Mhmm. Just have pi buttons. But the real math day on the calendar to me is today, this day, June 28.

And what makes June 28 the special math day is that, six and twenty eight are referred to as perfect numbers. Are we aware of this?

UnknownI never thought of six twenty eight as a perfect I knew what perfect numbers were, but I never associated those two together to make it a calendar day. Yeah. So Kayla and I have a history for this. We have a historical precedence, and it just happens that I took her to one of her first fish concerts on June 28. On this day. On this day. Kayla, was it your first show ever, or was it just one of your first shows?

UnknownI don't remember. You were young though. On this particular day, you were probably seven years old or something like that. Am I right? Particular day we're talking about? But it was a pretty long time ago. It was I wanted to say I mean, I have to go look on the Phish archives, but or for when they did this, but it was I think 2014, something like that. So Yeah. That checks out. Right? That would have made that would make you nine

if that's accurate. K. So that would make you nine on this particular day. And, you know, the way it goes when you go by the way, Phish is I have to tell people on this podcast, Phish is a band. A very special band that, I started taking Kayla to go see roughly when she was eight or nine years old. And on this particular day in Philly, the way it usually goes is you go, you take your you take your child to a show, maybe you leave a couple songs early, you don't stay for the whole show,

you know, so they ruin your show a little bit, but you take them home. And if they're awake, you know, you maybe talk about the show or not. Right? Usually, kids are fast asleep, you know, and you just drive home and that's that's that's that. But on this particular night, we I just brought up to you at nine years old. I said, you know what's a fun thing about today? It's June 28. Is that the number six and the number 28 are both I I had just read this, like, on Twitter. I'm pretty sure you frame this as

Unknowntoday is a perfect day. Yes. Well, I felt really good about taking you to official and you not totally ruining it. You actually had a good time. Right? Yeah. And I'm like, yeah. It's a perfect day. Let's talk about these numbers. Right? So this was like our drive home. Our drive home was so six is sick so what's a perfect number? Let's just, like, explain what it is. Right? It's basically the factors. The factors add up

to the number. So what does that mean? So if we take six, if you take one, two, and three, you exclude six obviously because if you add six, you're you ruin it. So if you take the factors not including the number one, two, and three, you add them together, you get six. Okay. That's kinda cool. Okay. Right? And then we did the same thing for 28. You get one, two, four,

which is 7. So 124714. 124714. 127. Sorry. 1 2 plus 4 plus 7 plus 14. That equals 28. Right? It's kinda neat too, but, like, who cares? Right? It's cool but who really cares that much? Right. What was fun about the drive home and what was meaningful to me and why I'm kind of like

miffed that we didn't talk about this last week in your when we talked about your origin story was that we spent the entire ride trial and erroring figuring out the next perfect number might be. Do you remember this? Yeah. I do. Right? We got into we got well in Good.

UnknownWe went through all the numbers and we're, like, trying to add up their factors. And then we were, like, circling back to, like, six and twenty eight. We were looking, like, what makes these numbers perfect? And then I think you went home and tried to create perfect numbers of the computer program. Well, so this is what ended up happening. By the way, this is like one in the morning on the way home from the McFish show. Perfect. Where I would have expected her to be asleep. And she's like, no. Let's try the next one. One zero two. What do we got?

One fifty one. Okay. That's not it. Like, we were it it it woke something up in me that made me realize that Kayla had energy, you know, she had energy for this And what we decided to do when we got home and again, this is like 2014 where, like, Raspberry Pis were kinda big. Like, turning them into actual computers was a big thing. So we went and ordered a Raspberry Pi,

and we decided we were gonna use its native its native operating system and its built in Python to, try to search for the next perfect number. Kayla, you recall this now? Yeah. Right? And we started writing a program to do it, and, you you know, we were it's like four I in range. You know? I don't remember how far out we decided we had to go, but at some point, the program blew up because we were looking for factors.

We were doing all kinds of memory intensive things that this little Raspberry Pi couldn't couldn't handle. Right? I'm surprised this thing didn't catch on fire.

UnknownRight? And I think after a couple of days of, like, waiting for it to run and think and all of this, I think we capitulated and looked it up on Wikipedia. Kayla, is this not right? Did we or did This is the correct story. Yeah. Right? But, like, this particular episode, I don't know. It woke something up for me about you, about, like, what we like, that this was unusual. We looked it up on Wikipedia, and it was, like, the next one was, like, a thousand something,

something in the thousands. And then the number after that was, like, the one of the biggest numbers I've seen, and I was like, oh my god. We would have never figured out

Unknownwhat the perfect number was. Like, was in the hundreds of thousands, I think. Yeah. Well Yeah. So the one that's after '28 is April. Oh, that's right. Yeah. And then after April, it's 8128. And those were the only four that were known in the pre computer era, basically. Really? Is there something with Mersenne primes here? There is something with Mersenne primes. Sorry. I I would say Yeah.

It's not pre computer, but, like, in the early days, I've I pulled up a little history, and these are the ones that the Pythagoreans knew of. Oh, the Pythagoreans. You talking to And the Pythagoreans whose motto was all those number thought that six was literally perfect, not a metaphor, but a metaphysical claim on reality to the point where they say six is a perfect number in of itself and not because God created all things in six days. Rather, the converse is true. God created all things in six days because the number is perfect. Now we know Like, they went they went deep to, the theological metaphysical claims on reality. Yeah.

UnknownThey were they were a bit of a testy bunch. Yes. They were. Right. We've talked a lot about the Pythagoreans. You know, that somehow the spirit of the Pythagoreans never dies either. Yeah. And so yeah. And then, a bit later on, Euclid proved that if two to the p minus one in prime, then two to the p minus one times two to the p minus one is a perfect number. He proved that. So he proved that. So for p two, three, four, five, six, seven, but I don't think it goes on forever

though. Because as of right, like to jump forward to today, I think there's only, like six fifty or 60 so perfect numbers that exist, that that are known to exist. But there's not really a proof one way or another if that's it or not. The largest known prime, I think it's a perfect number. It's somewhere in the ballpark of two to the 136,000,000.

UnknownTwo to the How it be and 30 prime if it has factors that if it has factors? Sorry. Yeah. Yeah. Yeah. It's if you have a I'm I'm looking at it right here. So you Oh, using the Mersenne primes. Using the Mersenne prime is a way to figure out which one would be the next, like, perfect. Got Got But it's not, like, an explicit formula yet. No. We do you think we can ever generate, like, all the perfect numbers?

UnknownThat's an interesting yeah. I don't know if we could. I mean, maybe we'll give it a couple more years with AI. I'm not sure if you've seen all of these proofs that have been in the past month or so that have been coming out using GPT 5.5. Maybe we'll learn a lot more soon enough. I mean, even a theorem like there are infinitely many perfect numbers. The way that we can say there are infinitely many primes, which I think Euclid

proved that as well. Right?

UnknownYes. Yeah. Euclid's perfect number recipe requires a a Mersenne prime to produce a perfect number. And are there infinitely many Mersenne primes, or do we not know that? I don't I don't think it's proven. Yeah. Very cool stuff. I wanted to I wanted to tell that story. And Yeah. You know, just again, like, the the journey of Magic Internet Math is, you

know, it didn't just start last December. Like, we've been we really have been building the lore around this for for a long time. Absolutely. I'm pretty sure I've celebrated Perfect Number Day on other podcasts in the past three years.

UnknownReally? Yes. But this is, yeah, special to be able to do it here on the Magic Internet Math podcast with you guys, especially with you here, Kayla. Right? Yeah. And Rob's kids are discovering perfect numbers right now as of this week. Isn't that amazing? That's right. That's right. Yeah. That's gonna be a Sunday. That was like the origin of perfect number day. It really was. Right, Kayla? That was like the first time we really started taking math seriously as a fun as like a fun family activity.

UnknownWhere it felt like exciting and like Yeah. It wasn't forced it wasn't forced out of some one of your classes you were forced to take. It wasn't you know what I mean? It wasn't, like Yeah. One of those things. Just natural. Yeah. That's where it really gets fun where it's like it's this is not for a class. This is this is just me being curious. Yeah. For sure. I have a book. I have a number theory book written by Shanks. Do do you know Shanks? Like, of, like, Shanks algorithm? Shanks algorithm is, I think, what

is the, is the one that says you only need two to the January, like, with the birthday paradox to Mhmm. Cover shot two fifty six. But Shank, I have a book by Shank's, and the first chapter is about perfect numbers. I haven't I haven't cracked it open. It's one of these like, this book is so old. I feel like the pages are gonna disintegrate if I engage too hard. But, it's been on my list. It's one of those on my list.

Anytime you get a book by, like, one of these guys, you gotta you have to get it. Yeah. Alright. Alright. So now I think we can move on

Unknownwith our little, perfect number preamble for celebrating the day, and we can start taking this to closing out the Elliptic Curve guide that I'm sure all of our listeners at home have been closely following along every step of the way. What chapter were we up to now, Brian? I wanna say chapter 11. Don't read too much into chapter 11. None of us are bankrupt yet. But Yeah.

UnknownI think chapter 11 is kind of, like, the last bit of new information as it relates to, everything. The previous ones were kind of just summarizing the parameters again. You know, I think Schnoor is also kind of where the real meat and potatoes are as it relates to the last, like, meaningful mathematical context as we care about for elliptic curve cryptography. And to reset the stage, not only for those at home, but also for Kayla,

Elliptic curve cryptography is the field of cryptography that basically Bitcoin relies on as it relates to the public private key link and the provability of ownership and being able to provide signatures in such a way that you can prove that you basically know a number that no one else knows. People, they use Bitcoin, they pay to public keys. And when they go to spend, they reveal private keys. And those private keys are what shows the link of I have the

associated of private key with the public key that you paid. And as another just refresh setup, I always refer to as if you walked into a post office and you see a wall PO boxes, all of those PO box numbers are basically public keys. And then the person with the private key is the one who can actually anyone in the world can send to a public key. You can send mail to any PO box address. But to be able to take the letters out, you have to have a mailbox key. Right? And instead of using fungible things like physics and lock mechanisms,

we use cryptography and math to show this asymmetry of knowing a number that no one else knows to be able to prove that ownership link. And everything we've gone up to up to this point has been about ECDSA, Elliptic Curve Digital Signature Algorithm. In an interesting bit of history, there are ways you can use the same kind of core mathematical principles in a more efficient way. When Bitcoin came out in 2008,

the competing and across the board objectively superior method of doing this is a snore signature,

Unknownbut it was patented. And the patent expired in 2010, but Satoshi came out with Bitcoin in 2008. So for the initial inception and launch of Bitcoin, we used ECDSA and probably one of the most famous, tech debt situations ever. We did not get around to upgrading to support Schnorr signatures until, 2021, which is when we did the Taproot upgrade. Taproot, as a very brief tangent,

did multiple things, of which snore signatures were one of them. And if you're a Bitcoin BIP aficionado at home, this is BIP three forty, three forty one, three forty two, And BIP three forty is the actual snore algorithm piece, whereas BIP three forty two is the new script type, moving to SegWit v one as opposed to v zero because we're computer people. We index on zero,

which gave you the new address type of Taproot addresses.

UnknownAnd It's because we're Python. We're Python people. I mean, right there are there are some languages that start on one. Correct? R, I noticed. There you go. R index is on one. And there's a couple other ones. And then three forty two is the actual validation of Tapscripts, the actual logic of them. But, the other piece of what Taproot is known for are masked Merkle abstract syntax trees, which are basically, you can have a Merkle tree, which I think we've briefly talked about previously,

where you can have many possible outcomes and you actually use a Merkle tree and you actually post the Merkle root to the blockchain. And that Merkle root allows you to unfurl up to two to the one twenty eight possible spending conditions. This allows you to have one single Bitcoin address and you can have up to 200, two to the 128 possible ways you can spend that Bitcoin address. And you only have to reveal, the spending information necessary and relevant to your address.

UnknownYeah. Like the needle in the haystack analogy. You know, like, you the Miracle tree is the haystack, but you get to post you get to show the needle. Yes. Right. And all you have to reveal to go down into the haystack, so to speak, are the tree of going down the Merkle tree. And you this is what's One called path. Right. The one path. And this is what's called the control block because you're controlling where within this plausibly in, two to the one twenty eight size haystack what exact sequence you're going to to get your exact condition.

And you commit to what your spending rules are ahead of time by hashing it. And this is what's called, like, a a tap leaf. Right? An individual leaf is the one given spending condition, and that gets hashed up into a Merkle tree where the Merkle root, the parent, the hashing of all the hashes that come together is the one that gets committed to on chain,

Unknownwhich is a very clever yeah. Would you say so, like, would you say Taproot was probably the the breakthrough in making Bitcoin programmable money? I think it's scalable programmable money is what it got to. Like like in a scalable sense because you can you can with prior to Taproot, you are able to have programmable money. Right? You can use lightning in a pre Taproot context, which uses up if to basically say either this condition or this condition. But it's inherently gated across

many reasons. And what Taproot allows you to do is one, they allow you to have up to a four megabyte actual size transaction. So if you wanna have an incredibly intricately involved UTXO that has many, many, many, many rules that can take up to four megabytes of opcodes and and pushing of data, you can do that now with Taproot, whereas you were constrained before with SegWit. And also, would say the other thing I think we're gonna get into now is specifically the Shneur signatures

are just in superior across every dimension. Right?

UnknownParticularly from privacy standpoint. Yeah. Yeah. So you've probably if you're in the industry of Bitcoin at all, may have heard the term MPC, multi party computation, right, where you have multiple computers or kind of you you would take a shard you'd have a secret, you shard it out across many parties, and then you combine them together to provide a signature. This uses a lot of very advanced cryptography and applied cryptography.

And from a security standpoint, it is a very common thing that those machine those systems get busted. Because there's multiple issues where you maybe have a theoretical abstract implementation that works, but in practice, you don't make it one to one to the theory. And then wherever those gaps are is where you can have money losing bugs or hacks or whatever have you. Right? And so Chnor in general, the way I like to say is, like, it's very easy to aggregate public keys in Chnor. It's such a trivial you just combine them together as opposed to having to do this crazy sharding scheme, which means you can have multiple parties in a multiparty computation system could each have their own public key. They each have their own public private key pair. And then all you have to do is just I give you my public key. You give me your public key. Kayla give me her public key, and we can just smush them all together and have one agreed upon public key, which is the combination of all of our keys,

which is Now from an abstract theory standpoint, way more straightforward than trying to do the sharding thing in a ECDSA MPC.

UnknownSo what I was gonna ask was that is it I think that it's the properties of the group that give this you know, that that make that possible. Right? Yeah. It's it's the way basically, it's the way a Schnorr signature or at least the variables in the short in Schnoor signature are Mhmm. Are aggregated. And as we talked about before we started recording, there's no division in Schnoor. That's it.

UnknownYeah. Right. So a Schnorr signature, not having that makes it infinitely easier to and not only just for the combining to make the public key, but when I take my signature, you take your signature, and we all take our signatures and put them together to make one final signature doesn't have division, which makes it infinitely easier to be able to securely and security wise understand what you're doing when you're creating a signature. So what it all means though and so this it has closure it has closure as a signature, which means that a signature plus a signature becomes a signature. And that's not true in ECDSA.

You can't that's it just doesn't it doesn't have that. In ECDSA, a point plus a point is a point on the curve, but this object called a signature doesn't have closure. And that's the power of Schnorr is that the signature plus a signature becomes a signature. And what does that mean for people or what would that mean if people were using it is that, you know, a two of three multisig wouldn't have to have three signatures where it's obviously where it's obvious that's what it is. Yep. A two of two multisig

makes it completely obvious that it's a opening of a Lightning channel. Mhmm. But under Schnorr, it would look like any other it would look like one signature.

UnknownThat's right. Yeah. And this is the privacy thing ultimately is that any threshold, which so the term we use in cryptography is a threshold signature where you have some sort of m of n, m required signatures for n possible keys. Any threshold that's frost. I'll get to that in a second. When you have a multi a proper multi signature in a cryptography cryptography sense, meaning a two of two, a three of three, an n of n. If everyone comes together, that's a multi signature.

That can just look like a single signature on chain. And that's what we I think we've briefly talked about it here and there. That's what MUSIG does. And from a privacy standpoint, it's great because you're actually totally obfuscating the security of how your Bitcoin is being managed because all of that computation is a multi party computation done off chain. And this is from a scaling perspective, why it's important is because you're able to do this math off chain, and you only have to post the aggregated final signature on chain.

Whereas if you were doing a two or three multi signature, you'd have to post two signatures to the chain every time. Right? And that's what's really by leveraging Schnorr, we're able to do this with MUSIG. The other one is Frost, which is for thresholds where you could have a m of n, a two of three, a five of, a three of five, a seven of 11, whatever threshold you want. And that could also just look like a single signature on chain. So you're totally able to make any plausible multi signature or threshold scheme for securing your Bitcoin look like just a single signature on chain, which is better for privacy,

Unknownas well as less data for scaling. Yeah, talk about that for a second. If you weren't here, I would talk about it, but you are the expert in wrench attacks. Right? Yeah. Anytime you Knowing spend Having which is saying from the perspective of, a wrench attacker knowing exactly what type of quorum your freaking keys are held in versus you know? Yeah. Speak to that for a second. Yeah. So anytime you spend Bitcoin, right, when you're spending Bitcoin, you have to do several things. One, when you pay earlier, said you pay to public keys. That's that's true, but it's not completely true. You also can pay to hashes.

And so the moment you do anything more involved than a single signature, you pay to what's called a hash, a script hat. You pay to a script. But since you don't post the script on the blockchain, what you do is you post a hash to the script. So the first thing you do when you spend Bitcoin, if it's anything beyond a single signature, what you need to do is you need to take the script and hash it and show that, hey. This hash that I'm trying to withdraw Bitcoin from matches this script hash I have here. That's the first satisfying condition.

Then you have to then execute a successful of the Bitcoin script stack. You have to successfully go through the stack and show it. And most times, I would say 999 out of a thousand, if not even higher percentage likelihood. You have a signature somewhere in there. The reason why you probably have a signature in there is because you don't want any person who knows the hash pre image of the script to be able to spend. You want something unique that only you know, which is a public private key pair. Right? And so when you go if someone from a surveillance standpoint knew who you were and let's say you didn't practice

good privacy, say, Brian, I paid you for, dinner one night, and you have your hot wall and then you you you sweep it to your cold storage and you post post it. And I I watch it for a year or two and see actually you're combining it with all of your life savings. When you go to spending, you're like, oh, I know for a fact Brian uses a two of three multisig, or I know he uses, you know, a three of five or whatever have you. Right? And the beauty of this is that the attacker, if you're using Schnorr and MuseSig and Frost,

all I know is that there's a signature operation's happening. I cannot from as an external party evaluate

Unknownwhat is that threshold, what are the circumstances around that. So from a private system You don't get an error message. You don't get an error message from Oh, two more signatures needed. Bingo. Right. Bingo. All I see is a pipe a public key, and then I see a private key. I see a signature for from those private keys. So it's great for plausible deniability for security, operational security, all of that stuff. And then so if I were to come with a $5 wrench to peek over the head, you could tell me like, oh, I have a seven of seven that's scattered all over the world, I can't prove you wrong.

UnknownThat's right. And you don't know if it's a passphrase or if it's a multi say. You don't know why it's not unlocking the key. Right. You don't have no idea why. Right. So, like, I mean, I think, like, it's not, like, gonna prevent this. There's always gonna be this first wave of wrench attacks, but Right. There's gonna be the I think in a world where in a world where snore was the standard, that that period of time would go go would go by very fast. The information of how futile those attacks are would get through pretty fast. Right.

UnknownYeah. Yeah. And then just some other just tail benefits of Schnewer. The verification is actually faster too. So when you're syncing the blockchain, like, this goes back to just, like, when you're looking at a signature, you're just looking at one signature. Right? And, you know, when you're like one that just, like, scales not just from the size on chain, the size to actually validate it. There's a couple small other quirks of there's no malleability

in in the Schnorr signature itself. Yeah. This is, you know, it's less relevant in SegWit and Tapper because we basically don't count the witness elements.

UnknownBut when you We covered this we covered this when we talked about quadratic residues. Yes. But we didn't mention this one thing that there are there are two possible there's always two possible answers. Right. Right. There's always two possible answers. And for The the elliptic curve equation is y squared equals something. Right. So there's always two possible answers. Two solutions. Right? And we we we handle this nicely with the well, one with SegWit in Bitcoin where we don't actually count the elements in the witness as part of what can determine the transaction. This is old Bitcoin history where

originally, Mount Gox blamed transaction malleability as to why they got hacked. And I would question if that's true or not. You'd have to take them at their word. I'm not sure if it was ever externally shown to be a reasonable way why they could have that. But the idea, though, is that you could actually have a given Bitcoin transaction in a pre SegWit world. You could get multiple transaction IDs. And this actually had to be fixed for Lightning

because if Lightning is a series of pre signed transactions, a UTXO, like, where your Bitcoin actually lives on the blockchain. The whole point of the blockchain is that you have an unspent transaction output saying this is my money and I can spend it in the future. You need to have, stability in the in the transaction ID. Once it's confirmed, it can't change. But the idea is that in flight, you could in theory have transaction ID a, then malleate the transaction, then get transaction ID b,

and you could have the and then transaction ID b only gets seen on chain. But if you were only looking at the unconfirmed transaction, you'd actually have a different one. But when you're doing this in Lightning, you're doing a bunch of off chain presigned transactions. And you need to have stability because if I have you and I have a Lightning channel, we keep paying each other back and forth, back and forth, back and forth. But then at, like, the second hop out of a thousand, I actually go up and I malleate the transaction, I go to close it, you would have no ability to contest that transaction because the chain is broken.

And so that's this was something that already needed to be solved for Lightning. And, you know, with in Shneur, this ties into like the I would say the last general point that we haven't touched upon is that, and we probably should just get reared in on to complain about this for two hours. Anytime. X Schnor, this was a this is not a requirement of Schnor, but Bitcoin's implementation of Schnor. We saved one byte on every key because we require that it's the x only key.

So we're always assuming an even y value when we, when we use a Schnorr signature in general. And so we always assume it is an even valued y values, and that kind of removes the the possibility of the y squared point he made earlier. There's always two possible answers. We are hard code assuming that it's always going to be the even value line number.

UnknownSo you're triggering something that I remember learning in in Lightning via whole it's okay to be odd, and this is why we use signal you signal odd numbers. Is is it related? Is that is that related here? Well, it's okay to be odd is just a general, like, software protocol spec thing that we do. It doesn't directly it's not directly related to this. But the idea is that if you have a protocol that allows for users to kind of upgrade and have different variable fields, what you do is you try to do is you try to have even values in the protocol be ones that are

ones that you really can't change and are really important. But then if I wanna peg some extra metadata or some extra information that the rest of the network doesn't have to care about, I I plug those in odd numbers. So it's a way to keep consensus across multiple actors. Not directly related to this, but it I I I see the parallel, though. Yeah. And then the I guess final thing is just cleaner security proofs, which we like. We we like having less assumptions, less, abstract math required to be able to prove that things are secure.

UnknownYeah. It's, I I think it's a little bit of a shame that what Taproot's sort of known for are, you know, quantum vulnerability and Mhmm. JPEGs. Right? Unfortunately, that seems to be that's like what the revisionist history would suggest that Taproot that what Taproot is meant to make us quantum vulnerable. Or maybe it was the quantum vulnerability is may is I've heard that as to why a lot of devs didn't build didn't really build on it. Yeah. I think, starting today, people would probably start saying that. I don't think quantum five, six years ago was really the relevant thing. There's just a certain stickiness with Bitcoin that, like, if you have code and and the code in Bitcoin,

the code is the money. The code is the money. Right? So if you have a engine and a flywheel and it works, why touch it? Like, is why Coinbase didn't allow you to even send to a Taproot address. From from their wallet, they didn't even support sending to a Taproot wallet until maybe a year and a half ago. Yeah. That's And then they they only over the ETF. That was that was like embarrassing over the ETF. Yeah. And Bitwise basically beat them up to do both those things, which is great because now it's done. When I first started Anchor Watch, one of the reasons why I did not turn on Taproot is that if I had a customer who was at Coinbase

and they were gonna move their funds over to me, they would not be able to, which is like, you could say what you want about, you know, poo pooing Coinbase and their product road map and what they focus on, but it's the largest exchange in the world. And if you're saying that I'm gonna be Taproot only, I am explicitly saying that the largest exchange in the world can't interface with my wallet. That's a rough break. Right? So those problems have now been solved. We're working on Taproot right now at Anchor Watch. It's one of these things, though, that was very relevant for being able to, like, get over that hump. And so

we're this is why I talk about, like, the largest tech debt this you know, thing ever is that because of that patent, Satoshi couldn't start with Shneur out of the gate. Yeah.

UnknownYeah. And and but it's sort of now viewed because we still don't we still don't have Schnoor. Right? We have Taproot. We have these other externalities. We don't have Schnoor yet, really. We tech so you technically can. It's just not widely supported. It's not a standard. It's not we don't have it as a standard practice saving the world from bad privacy. Right? Yeah. So there's a music BIP now, BIP three twenty seven. It originally got a BIP number in 2022,

so, like, a year after. I think it's actually an interesting thing. I'm gonna look it up right now. So there's something called RFC 60 RFC sixty nine sixty seven, or is it sixty seven sixty nine? It's sixty seven sixty nine. Sixty seven sixty nine, which is, wait. Deterministic nonce RFC. ECDSA sixty nine sixty seven

is you wanna always so when you're doing it like a multiparty computation well, in general, when you're signing a Bitcoin, we've definitely talked about this before. If the nonce gets reused multiple times for a single public key, you actually leak the private key because you are actually able to trivially do the division and work your way backwards. Yes. Literally y equals m x plus b math can Yeah. Like, it's it's Yeah. Not Yeah. You basically you've you've you've filled in enough of the, variables where you trivially can just do some quick division and get the private key. So there's a spec for how you always have deterministic nonsense

with ECDSA. I think last I checked, there actually isn't a formal spec. It's been what's kinda hacked together for deterministic balances for ECDSA. And that's, like, a little bit of that's not a complicated thing to solve, but it's a relevant thing to try and go figure out. Chapter 12,

Unknownthough, in the chapter 12 in the study guide does, you know, lay this out. Yes. Yes. I'm looking I definitely had a tweet about it. Snoring. Do you wanna drop your snore factoid? Drop my snore factoid? Oh, yeah. You mean the fact that snore there's a Yiddish term called snorer? Yes.

It was brought to my attention by, by a friend that, schnoor there is there's a word there's a Yiddish term called schnoor, which is actually a, you know, an over it's funny that they had to create a Yiddish term for an overly transactional human being. But, so this whoever this was made for must have been just so over the top unbearable that they had to create this word called schnoor. And this person was asking me if if that was the origin of

of the term schnoor signature. But, actually, no. It was a man named Klaus Schnoor Mhmm. A cryptographer, I think, who is a he lives in our time. Correct?

UnknownI believe so. I mean, he had the patent he had the patent in Bitcoin back, you know, I'm looking to see I've seen it right here. Yeah. It's Pausch Nor. 1989. He filed He died he died last year. Oh my Wow. He died a year ago. June 8. Oh, man. Right before perfect number day. Right. He didn't make it. Oh, man. Yeah. So but it does so it says here he filed the patent in 1989,

and it expired in February 2008. But I guess it was still too raw.

UnknownI think yeah. Like, was a little too raw. I mean, this is the other thing too is that you wanna battle test systems before you put the whole money on it. Yeah. Right? So Bitcoin used the following sixteen years to make sure it was totally good and then turned it on. Yeah. So, like, there's there's a whole ecosystem around Bitcoin to get this. Like, you technically can do it. There is a bit for music. Right? There the thing, though, is that you do want there to be a whole ecosystem. Right? So right now, like, our back end and Anchor Watch is built on Rust Bitcoin.

There really isn't a great way to do music keys within Rust Bitcoin yet. It's something I I've I've kicked the tires around, and I think it's something that we're getting to a point where everything else is so rock solid stable. Like, the work can start being done on that. Music has now been added to the libsec p two fifty six k one library, which is a big milestone. So, like, we now have, like, the core foundational of the cryptographers have a designated spec, which once you get that, you get a really big unlock is that you can actually have the hardware wallets, which all just import this code.

Like, every hardware wallet is not writing their own let me take that back. No reputable hardware wallet is rolling their own cryptography. Right? They are taking the standardized spec library and then implementing it. So once it was added this past year to music was added, you now have, like, the cold card which supports it, and you have the ledger which supports it. Right? So, like, you start getting, like, the initial, like, green shoots of an ecosystem, but now you need to have wallet, providers to be able to have the infrastructure, be able to do this combining of the keys securely and everything. Right?

So, that's all moving in the right direction. I think we're we're we're now starting to get there. There's now been enough groundwork laid out to make it worthwhile.

UnknownDo we wanna cover the current events at all around If you'd like. I think we should since we're here. Right? So Yeah. I mean, also, just as we we've clear we've closed out the study guide now. Right? We could put a pin in the Elliptic Curve study guide. Have we finished that? I wanted I I I do tend to like to read out loud some of these Steiner blips. Sure. And I do Do wanna about about the Steiner stuff that was being said here.

UnknownSo there are these little boxes called Steiner lenses where my clanker did it. My clanker, I asked when he created these study guides to, you know, tell us what Steiner would actually think of a lot of these concepts. Totally wanna Do Do we wanna do that to close out the study guide or do you wanna like do that to close out the study guide next episode? Like, I just wanna make sure we've now gotten to the end of the study guide is kind of what I'm getting at. We have. We have. Congratulations

everybody for sticking with us. Like 10 episodes, not too bad. Something like that. And there are exercises that you should all go to the end and do because they're fun. Yeah.

UnknownAnd helpful. We can go right to topics of the day if you want and we can I think this, yeah, let's do that? Let's let's then let's acknowledge the study guide complete right now. Yeah. And let's let's but let's and let's, hit this current event because it's all kinda related. Right? I mean, almost everything that we just talked about happened because a single person was maintaining, you know, was largely main he was developing music and maintaining the LibSec library.

Oh, yeah.

UnknownRight? Yes. That's the current event I'm talking about. Yes. Yes. Yes. I don't know what you had in mind. So no. No. No. I wasn't sure where you were going. Yeah. So it looks like Jonas Nick was stepping down from the, I would maintain a role. I'm not sure if they formally have lead maintainers in the libsecp library, but Jonas is definitely, up there with the heavyweights of, you know, the actual work around the libsecp library and standardizing it. Brian and I have met him on several occasions,

including at OpNext or back in April, was it? Got to chat with him a little bit. He's great. He's now the director of research at Blockstream, so he's managing and spinning many, many plates, and kind of maintaining all of the commercial work that Blockstream is doing. Future guest. For sure. Future guest. Yeah. He's he's he's in. He just wants us to know a time and a place and a topic.

UnknownWe need to delay all the scaffolding out though. Yeah. We needed to do all this. And I think that just again, if we talk about, one of the slides I had in my in my, keynote, you know, for, in Prague Mhmm. Around all of this was also that, like, don't overestimate don't overestimate Bitcoin Core and not see that it fits that it's in it's in a ecosystem Mhmm. That we need to worry about equally a lot of other things. Cool. This is a great point is that everyone thinks about Bitcoin,

Unknownthe Bitcoin core organization as Bitcoin slash Bitcoin on GitHub. Yeah. But there's multiple libraries with it. Yeah. Like like, well, if you go to GitHub, right, .com/bitcoin, if you're looking at the code base, it's Bitcoin slash it's Bitcoin github.com/bitcoin/bitcoin. The reason why is the actual node implementation is part of the Bitcoin project and the Bitcoin repo, but there are other repositories that are really important for how Bitcoin works within there. Right? First, you could think of, like,

the BIPs repo. Mhmm. You know? Then you also have, you you have you have libsec p, which is the cryptography one we're talking about, which, like, is, I would say, really the foundational library that everything on Bitcoin gets built on top of. If if there's a bug in the in the libs the sec p 256K1 library, the rest of it doesn't really matter. Your node doesn't really matter if the cryptography breaks, which is kinda going to, like, the quantum discussion. Right? They're like all of the software around it doesn't really matter if you can't uniquely hold a private key to yourself.

UnknownYeah. And it's also like, you all know my story. Like, I I had a I had a great time looking at the Bitcoin repo. Formulas in English, everything I knew was supposed to be there was there, and it all made sense. But, like, that libsecp library, I think, is really hard. It's I'd I'd say for a layman, forget about it. You're not you're not gonna walk away from there with any good feelings. And if you've if you are somebody with some mathematical aptitude who's maybe had a career

that's mathematically related, they're somebody like me is gonna also gonna walk away, like, actually really upset, not knowing what the hell to do. And it just make what what it does is it makes you realize that those people are far I think far more important to at least in the stage that we're in right now. Right? The 2026 status of things. I think they're more important to, to what that library does than the Bitcoin people are to what to what Bitcoiners need at a, like, to what we need at a Bitcoin.

Yeah. Right? So I think that that's one of the reasons we do this podcast actually is to shine the light on the LipSec the LipSec p library Yeah. And get us a little bit more so that we're less dependent on them when some type of crisis happens.

UnknownYeah. Absolutely. And there are other people who work on that. I'm actually looking over the past year. The you know, you have Tim Ruffing, real random. You also have, Habosto, who's a Bitcoin core maintainer, who I've met this past year, you know, working on this stuff. Everyone wants to care about the little drama inflection points, but this is genuinely where the real work is being done. And and to really just observe

how rock solid and how small and incremental, the code is in the libsec p library. I'm looking over the past year. Hibasto has the most commits. He added 484 lines and removed 435

Unknownlines. Wait. A year of work. Wait. Four eighty four was not one of the that wasn't the perfect number. Right? What what was the third? What was the three digit perfect number? 496. Was it 628 or hold on. I have it right here. One second. It was it wasn't 625, was it? No. No. Nope. It was it was 496. 496. So close. Oh, he's, he's 12 commits away. Well, he's 40 scaring the shit out of us. Yeah. That's right. Right. Right. Right. Right.

He, but, you know, over a year of work and on net, he added 40 lines of code. You know what I mean? Like, this is, like Right. The level of precision we're at. A year of work and you added 40 lines of code to the repo. Like, that is the level of, like, very tightly dialed in stuff here. And so, you know, it's it's critically important stuff, but that was just a little, I I messaged Brian when I saw that happening was kind of like a so the other two maintainers at the moment are

Tim Ruffink, who I mentioned, and Peter Willa, who I think we've only mentioned a thousand times, on this podcast.

UnknownWhere's Andrew on that spectrum? Has I'm looking over the past year. So Andrew's definitely I don't think Andrew's done any contributions in the past year. I can look at all time and probably find him, though. Yeah. Yep. Polstra stopped committing to libsec in 2021, like, just after Taproot. So he wasn't on the the core stuff. I mean, I'm I'm now looking at the all time the all time greatest hits. And Sipa is the number one committer with 477

commits. He added a 103,000 lines and removed 86,000 lines. Now mind you, his first commit was in March of in January 2023. Sorry, January 2013. Wow. So you're talking thirteen years of work and he added like 20,000 lines of code over

Unknownfifteen, sixteen years. Now the stakes are high. I would would you say Taproot raised the stakes on SecP two fifty six k one? Yes. And so far as when you start the reason why I would say raising the stakes as a yes is one, we had this, like, ECDSA algorithm that we were all, you know, we're all really happy with and using. So to whenever you're upgrading Bitcoin, you are changing the engine on the plane while it's 30,000 feet in the sky.

So, like, you really can't mess that up. Right? So in that sense, yes. Also, the nature of the aggregation of signatures, I would say, are you're you you need to be really sure about that because you don't have this genuine this original assumption now where each signature could be self contained to a single key. You're combining all of these and making sure that's all done correctly.

UnknownSo I I can see that. Yeah. So if I'm gonna say, if we're explaining like, if we're explaining the state of things to the world, we might say that the important the most important aspects of Taproot should have been expected to move slowly, Mhmm. Very slowly. And Yep. It it's not surprising that the thing that no one likes that was exploited happened right away because it's it was a bunch of reckless people who didn't care how it turned out. Yep. And, you know, it came. That, you know, that came and went. But, like, the reality is the what we wanted at a taproot, which probably

began with Segwit, right? This arc was always going to be very slow. Yeah. And maybe this fear of quantum also made it even slower potentially. But I think it's just people should have this context that there is a reason why this exists. Right? Taproot was not put in to strictly for people to abuse the segway discount and size limit.

That's not the reason it was put in. It's just that you're not seeing it's gonna be a long time before we really see we before we really see it. And it's okay to have a low time preference for development. Right?

UnknownYes. I also do wanna call out, Jonas in his replacement for maintainer suggested the stack, Sebastian is his name, to become maintainer. And I just pulled up his Twitter profile, and he follows me. So we can get him on. Hell yeah. Yeah. We should grab him. And I'm also realizing yet, you haven't No. I haven't DMed him yet, but I can I can take that off? But I'm realizing who we should actually have on. We should we should try and get Peter Willa on, to just talk about libsec p because he is the one who really started that. And we've talked about it before, but the work of Satoshi having a couple 100 lines of code, which was the cryptography in the original thing, and slowly realizing, wait a second, this is really important. We should be

digging way deeper and optimizing this. I think that that is probably our that would be the ultimate capstone magnum opus of the ECC work where we could also, it would be funny too because whenever anyone wants to have Peter on, they wanna talk about quantum. They wanna talk about a thousand other things. We'd be like, no, let's just talk about the past fifteen years of work you've done taking Satoshi's 300 lines of code and making it to the LibSec p library today, which is like fifty, sixty thousand lines of code.

UnknownYeah. And I I would love to ask these guys if it burns if does it burn them out to work on this the way they do being the only ones that understand it. Yeah. I would imagine that they don't like being this priest class. I would imagine that. Basically a road to burnout, and they can't really do the things they wanna do. So Andrew's been working on his things, and he's not so he stopped contributing to the library. Right?

And you just yeah. Like, I would love to hear that because that should concern people. Like, this to me, this is the kind of thing that should concern you as a Bitcoiner. Oh, Andrew not, contributing anymore? Not so much Andrew not contributing, but just the general fact that there's so few of these people, and they're so prone to burn out potentially from the pressure and being the only ones and how important it is and how, you know, maybe really what we're doing is like the thing that needs to be done because maybe what we do is we start creating some people that are capable of eventually

Unknownlending it lending a hand on this. Yeah. I mean, that'd be a good thing to talk about the ecosystem of developing talent because if people say it takes five years of being a developer before you could really, like, cut your teeth on the Bitcoin software, That was that a decade for limsec p? Yeah. Jeez. When you say so that's just being general. If if you're I'm a developer. I'm a Bitcoin like, I'm I'm a developer who is able to write software. Mhmm. It's like a five year it's like a five year journey to go from I'm a generic developer to I am a developer who can push meaningful contributions forward in the Bitcoin code base.

Was that how it works for you? Five years Well, I've never I'm I'm not a Bitcoin core contributor.

UnknownRight? What about somebody like Evan? Somebody like Evan. Oh, Evan could. Yeah. Like and I I probably could now do small things because I'm so in steeped in how Bitcoin works on the application layer. I'm not walking in like a random developer who's never touched Bitcoin before. Right? Like, I've you could probably cut mine down to, like, two years or something. But if I'm working specifically on c plus plus code in the node software to try and do changes and do things, Maybe with AI, you could cut it down a little bit, but you don't wanna really be,

slot you don't wanna be inconsiderate and start slopping up PRs to have a review. But I mean, yeah, you wonder if that's gonna be a thing. Maybe. And the most like, this is a common thing that's said in Bitcoin. The most constrained resource is actually just review hours. You have to court someone to take time out of their day unpaid to go sit down and review code that you wrote. Right? Yeah. And that is like because and you need reputable you need reputable people reviewing your code and changes before it can be seriously considered.

So you need to kind of, like, build your own social capital in time. And what usually that starts with? You start doing your own review. Right? You start you start reviewing other PRs. You just start trying to, you know, value for value starts trying to contribute before you start taking.

UnknownRight. That's it's a whole like, open source being in the open source ecosystem is its own its own thing. Absolutely. Yeah. So I I like I said, look, I think that I'm glad to put more attention on sec p two fifty six, and I think people should be at least viewing that on equal footing. If anything, I think it should be getting more attention because less people understand it and it's far more consequential.

And to me, Bitcoin Core is just maintaining anyway something that it it like, you also have private industry that Mhmm. Will develop things that it wants.

UnknownRight. Right? Again, it's gonna take a little time. Yeah. And I would even say that libsec p in its proper place should be a commons. You don't want Yes. Oh, you know, Acme Inc being the one who runs the security for the entire protocol. Yeah. I thought you were gonna say MicroStrategy. So Sure. But they would have to yeah. But they're they're pretty clear on not wanting to give money to developers, period. Right? But good news. BlackRock has agreed to take over the cryptography. Well, this is this is our our mutual friend, Shinobi, talks about a catastrophic end state where all of the important work in Bitcoin is done by a handful of companies who have a big Bitcoin bag.

And then that is not a way to run an open protocol, especially talking about the security of these systems. It's true. I mean, I remember when the ETFs were coming out having a strong

Unknownview. I don't I I if I review this, I don't know if I'm right about it anymore, but I remember having a strong view that I didn't like the virtue signaling of, like, a Van Eck saying I'm gonna contribute, you know, 2% of our our profits on the ETF to open source development. For some reason, I that rubbed me the wrong way at the time. I haven't really thought about it too deeply since. I just remember that rubbing me the wrong way, like, oh my god. Are we gonna end up, like, just having these ETF providers dominate

the development ecosystem now because that's where the money's coming from? But then it I know. Maybe another part of me is like, well, anyone else is able to step up and fund these people too. So Right. Maybe we shouldn't be too uppity about where the money comes from. Yeah.

UnknownWell, the there's I and I'm realizing yeah. This is this has been a good conversation. I'm just realizing now that we should definitely just have Peter on. Because he's the actual if there was a lead maintainer, he's the one who took it and started the initial repo and started pulling all these things. And I think that, like, I don't think I don't think we have to do it as the next episode. I think we can let this kind of marinade for a bit. We can have we decided what our next major topic is? Is it hashes based off Farrington's decree? It's I think so. At least we're gonna start there. I think I started making the document. Yeah. I started building the document.

I've been editing a book called Every Byte Tells a Story. Okay. And that's gonna be available at Magic Internet Math, and that's gonna basically give you the ability to do everything Rob said you should be able to do during this episode, which is, you know, see how multi you know, you wanna know how op check multisig pushes op codes to the sack to validate you know, to actually validate a signature to throw a zero or one. This book will give you all of the tools to be able to do all of those things to know what he's talking about. Mhmm. What pay to pub key versus pay to pub key hash versus

you know, where does OPIF happen in Lightning? And what is an HTLC? Like, this is gonna be like a complete encyclopedic. It's gonna be like an encyclopedia of essentially, hex bytes on chain Yep. For for all things. And so I'm basically done editing that, and I'm gonna actually move that to publishing. And now I can focus on the hash functions document,

Unknownwhich I've made two or three chapters now. I think we could even even before that book is done, we could kind of, at a high level, just start the conversation on the hash functions. And genuinely after hashes and ease one, I think hashes are more contained of a system than elliptic curve. We definitely show we definitely we definitely started with our vegetables in this podcast. We started we went right to modular arithmetic, and we just went right down to the the You call those vegetables? If the the that's like, those are like brussels sprouts cooked in bacon fat. That's right. With honey drizzled on it. Right? Yeah. Yeah.

And then from the sake of, like, the ground truth of Bitcoin, like, this is, like, this is the roughage. Like, you have to understand this stuff. And after that and hash functions, I think there's a whole universe of things we can start talking about. And we could probably kick off a hash conversation, pretty low stakes, and just starting to lay it out, figure out from there where we wanna take it, and while you get the rest of the study guide stood up.

UnknownYeah. I like the direction of going in there. No pun intended. I like the direction of going into hash functions mostly because I think that it's one of those it's just the journey there is like, do we we understand really what they are and can we figure out if others who we would hope understand it actually do? Like, that's really to me the interesting question of why to dive in. Yeah. Well, once a hash function is kind of like observed to be secure, everyone's like, oh, then you hash it. Right? And then like,

there's a there's a huge leap there. And unlike, several things. One, because it's deterministic, I can run on my computer. You can run on your computer. We can make sure we get the same answer for alignment. It's not like I can't you and I can't run the same ECTSA signature because I'm not gonna give you my private key. Right? So it's something that's easier to kinda, like, glaze over because, like, once you have the software and the implementation, you do all of the checks and it works, you're good. Might be fun to get into a little bit of the the theoretical principles around these more too for a better understanding of the system. Yeah. Absolutely.

I I don't know if it's as clean as the group theory stuff. So it might be a little it might be I don't think it will be. I don't think it's gonna be as clean because it doesn't it necessarily doesn't have to be as clean. And it's kind of and maybe as a little teaser, you never wanna be you need to have very clean tight rules because you never wanna be able to have my private key and your private key result to the same public key. We never wanna have my private key signing something be able to sign for something in your transaction data. Right? You never want those things. So it has to be very, very black and white clear. Whereas

security of a hash function just kinda talks about the search space where, okay, if I have a thing and you have a thing and then we get the same hash, it's called a collision. And, you know, like, it's hard to do that. We could start getting to I know Google, had a paper about the collisions for SHA-one. We could talk about Peter's hash pinatas, which I know we've mentioned before. Peter Todd's hash pinatas, we can talk about that a little bit. But everything doesn't immediately fall apart and break is the is the key point. That's right. So like it's a there's there's definitely a little bit more wiggle room for hash collisions compared to private keys breaking.

UnknownNice. So then the let the the final note, just because I wanna make the note. Yeah. It's something that Kail and I discussed last week, but if you're doing this series and you didn't you didn't catch that, I wanted to make one note for completion, which is that, we've never we're gonna finish this now, but we've never actually proven that the points on the elliptic curve are a group. We Yeah. We were talking about that last time. Oh. So we took it for granted. That's a lot more advanced.

Thus right. So thus far, that's being taken for granted, and we're trusting math books Oh. On that. Fun. And so if you wanna know more about that, go listen to the episode that we just that I just published that Kayla and I did. Excellent. It was like we just just I I figured it out. I'm like, oh, no.

UnknownWe can go one layer deeper. Yeah. Thought we were done, and that's how that's how insidious this can all be. Yeah. If you really wanna make sure you're validating everything which Absolutely. Unfortunately, that's the agreement we all made. You know? Yeah. And so if we are gonna be true to that and not LARPs, Yeah. We kinda we're gonna have to we're gonna do that proof for you. At least, like, skim the paper and see. Yeah. So I basically assigned Kayla. I said, Kayla, this is your first assignment in the Magic Internet Math Academy is to do this proof for us.

And at some point, walk us through it.

UnknownThat's a cool thing because I'm I'm thinking about because the project I'm doing right now is that I have to basically narrate a story about a mathematical breakthrough. And I have to trace back its origins to really understand the big picture. And maybe maybe that's what it takes for the elliptic curve being the group. Yeah. Are are you so you're do you wanna talk a little bit about your research before we close

UnknownOh, since you're okay. So I have to tell, basically all this story culminates and there's this topic in a number theory that has been studied since Gauss, so it's like a two hundred year train of thought. And in like 2014, this PhD came out with his thesis that basically says, we can visualize all of this with a cube. And so we have to trace back like,

basically, we they were studying binary quadratic forms, which is like a special polynomial, like a homogeneous polynomial. And people found out that I think they found an isomorphism between binary quadratic forms and the class group. And the class group measures when unique factorization fails, which was created because studying, like, integer rings of quadratic fields, people found out that the golden rule of

you can break every prime number down into smaller no. You can break every big number down smaller pieces uniquely. That actually breaks when you get, like, more advanced. You add, like, square roots in there. And for us, by the way, without We have to figure out, like, why did this fail? Why does unique factorization fail? And how far away is it from actually being a unique factorization domain?

UnknownSo, yeah, we talked about here for us. Without unique factorization, you don't have a finite field. And you can't trace back, like, what the number's origins are if there's, like, two or more different ways to break it down. So interesting. So you're gonna get to do that. And then as a template, you're gonna you're gonna prove you're gonna do this proof, and that's gonna be another potential discovery. Like, you know, that's I think that

I think it's just as significant. When you when you go ahead and prove that the, points in the elliptic curve form a group, it's gonna be just as significant as tracing

UnknownTracing back, like, what does like, what is the actual study? What are we actually visualizing? Yes. And that's why also I stress math history a lot. I think you wanna understand the biographies of Gauss and Euler and all of these guys because you're now doing you're one of these people right now, and you don't have a good sense for that. Yeah. I think it's good to have these figures who, again, like, they they were just genuinely curious and wanted to solve these problems.

Awesome. And they, like, went above and beyond and contributed to, like, every field.

UnknownNice. Anything else? No. Not that on my end. Alright. Magic Internet Math Elliptic Curves study guide is in the books.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Exponential Got Less Exponential

    Nov 2, 20261 hr 10 minAudio

    In this episode, we picked back up on our red-team tour of Bitcoin cryptography and tried to make sense of the last few weeks of fast-moving math news. We unpacked why the recent excitement around the exponential time...

    Credit the people on this episode

  2. Red Teaming Elliptic Curve Cryptography

    Oct 11, 20261 hr 27 minAudio

    OpenAI just dropped roughly 300 math papers with Lean proofs, mathematicians are crashing out, and the vague-posting has started: "math is breaking, so cryptography is next." Brian and Rob Hamilton refuse to hand-wave....

    Credit the people on this episode

  3. Hash Functions, Linear Algebra, and Alice in Wonderland

    Sep 28, 202650 minAudio

    Magic Internet Math: https://www.magicinternetmath.com/ Shout out to https://x.com/uncleJim21 for the Alice in Wonderland story and the following rebuke of Lewis Carroll: https://www.youtube.com/shorts/33FYXpEWgyA This...

    Credit the people on this episode

  4. Kayfabe Math is Now Mainstream

    Sep 13, 20261 hr 17 minAudio

    We’re back because math keeps refusing to stay in the classroom. In this episode, I unpack two very different kinds of “math pop culture.” First, we revisit entropy, seed generation, and wallet security through a...

    Credit the people on this episode

  5. Legalize MATH

    Sep 2, 20262 hr 2 minAudio

    Magic Internet Math is back. After the summer that set the Bitcoin world on fire - the Cold Card entropy disaster - Brian and Rob Hamilton (AnchorWatch) reunite for a two-hour deep dive on the event Rob spent three...

    Credit the people on this episode