S17 E19: Amaury Séchet on Forking Bitcoin Twice, Quantum FUD & Scars from the Block Size Wars

Two years after our first interview, I decided to invite Amaury Séchet(@deadalnix) to find out what the Greg Maxwell of big blockers has been up to lately. This interview turned out to become a deposition, a history seminar, and a live audit of everything the big block movement got right and catastrophically wrong, delivered by the man who ignited it.

In case you’re new here, you must know that Amaury is the developer who forked Bitcoin in August 2017 to create Bitcoin Cash, and then, three years later, forked Bitcoin Cash itself after the funding dispute that ended his tenure as its self-described “benevolent dictator.” That second chain became eCash (XEC), a Bitcoin descendant that keeps Nakamoto proof of work untouched and bolts on a second consensus layer, Avalanche, for near-instant finality. His second appearance on the show ran four and a half hours, and we used the time to do something I had wanted to do for years: walk through the entire fossil record of Bitcoin forks with the person responsible for two of its five major branches, while he corrected my chart in real time.

I opened by calling him the Greg Maxwell of the big blockers, the technical authority to whom an entire movement deferred, which is precisely why the politics around him turned so vicious. He took it with a shrug: “Okay, fair enough.” Then he went to grab a drink, and we began.

First, a Disambiguation: How Many Ecashes Are There?

Because this episode keeps colliding with the question, and because the naming situation is about to get confusing again, here is the short genealogy. The word “ecash” predates Bitcoin by a quarter century. In 1983, David Chaum published Blind Signatures for Untraceable Payments, the founding document of privacy-preserving electronic money. His company DigiCash, founded in 1990, sold a product literally called “ecash.” Microsoft reportedly offered $100 million to bundle it with Windows 95; Chaum refused; DigiCash went bankrupt in 1998. The cypherpunk mailing list spent the next decade trying to rebuild the idea without a central issuer, and when Satoshi Nakamoto circulated an early draft of his whitepaper in 2008, the file was allegedly named ecash.pdf, titled “Electronic Cash Without a Trusted Third Party.” The published version kept the phrase “electronic cash” and added “peer-to-peer.”

Amaury’s fork, launched in November 2020 as the BCHA side of the Bitcoin Cash split and rebranded to eCash (XEC) in mid-2021, reached for that same lineage, and its marketing explicitly invokes Milton Friedman‘s 1999 interview about “ecash” as the missing link for financial freedom on the internet. In 2022, Cashu and later Fedimintrevived Chaumian mints on top of Lightning, also calling themselves ecash.

And in 2025, Paul Sztorc bought the ecash.com domain and announced that his BIP300 drivechain hard fork of Bitcoin, launching this very month at block 964,000, would simply be called Ecash (the full story is in S17 E21 of this podcast). One history blog aligned with Sztorc’s project argues the term is a forty-year-old piece of cypherpunk commons that no one can trademark: a tradition, not a brand. On the evidence, that is hard to dispute.

The coincidence produces one of this episode’s quiet ironies. Amaury explains his own naming decision as a confluence of things: it is a fantastic name, it was Satoshi’s working title, and the e.cash domain happened to be obtainable. The owners of ecash.com, by contrast, wanted “an absurd amount of money” back then, so he passed. Years later, Sztorc apparently paid it. And as you will see below, Amaury’s critique of drivechain withdrawal times lands directly on the design Sztorc is about to ship. The two Ecashes disagree about architecture while claiming the same patron saints, Friedman and Satoshi’s filename. The market will sort them out, which is exactly how cypherpunks say it should work.

The Price Question, or: Network Effects Are a Moat

I opened with the rude question. Since our last interview, XEC went down while BCH went sideways to up. Why? His answer has no copium in it: network effects. Bitcoin Cash was big enough, early enough, that it is integrated into everything: every exchange, every trading pair, every liquidity pool. He is “actually kind of disappointed that it’s not doing better.” eCash has to rebuild all of that from scratch, in a market that is far more crowded than 2017 and in which the very idea of digital cash has faded from fashion. The past few years belonged to memecoins.

That sent us into a shared rant about casino nihilism. When the market expects everything to pump and dump within six months, Amaury argues, “making a good project is actually a liability.” Why spend a year engineering something real when you could ship five pieces of garbage in the same time, since they will all be dead by Christmas anyway? His one hope: bear markets clean this up. Eventually.

Quantum Computing: Unsolved Physics, Oversold Timelines, and the Confiscation Creep

The week we recorded, Bitcoin’s discourse had cycled from covenants to spam filtering to a full quantum panic, complete with BIP 360 (a soft fork proposal that would effectively freeze Satoshi Nakamoto’s coins before a quantum computer can take them) and a counter-proposal launched the same day. Amaury, whom I crowned the king of hard forks before asking, dismantled the panic in layers.

First, the physics is genuinely unsettled. Quantum computers rely on the collapse of the wave function upon observation, and nobody has a rigorous definition of what “observation” means at scale. Depending on the answer, large-scale quantum computation may not even be physically possible. He has a physicist friend who worked in the quantum industry and half-hoped the machines would fail, because failure would mean new physics. That was ten years ago, and the question remains open.

Second, Google’s much-discussed announcement giving cryptography until roughly 2029 looks, to him, like marketing. A quantum computer that breaks elliptic curves also breaks HTTPS, SSH, and the signature infrastructure of Google, Facebook, Amazon, every bank, and every government. Crypto would be “not even remotely close to the main problem.” But crypto is the only target you can publicly threaten to break without legal consequences, and it is a community that loves controversy, which makes it the perfect vehicle for getting people to talk about your quantum division. I added the obvious corollary: announce you can break a bank’s cryptography and men in suits visit you. Announce you can break Bitcoin and you get a Wired profile.

Third, Satoshi’s coins are the ideal bounty precisely because there is no claimant, and because the oldest coins pay directly to public keys (P2PK), with no hash protecting the key, making them the most quantum-vulnerable UTXOs on the chain. His imagined scenario deserves to be quoted: Satoshi returns, says “Just leave me alone. Leave my coins alone,” consolidates everything into a single modern UTXO, and dares the quantum crowd to try.

But then came the turn that gives this segment its real weight. The thing that actually frightens Amaury is not the quantum computer. It is the normalization of confiscation. Ten years ago, proposing to freeze Satoshi’s coins would have gotten you laughed out of the room. Now it is a BIP.

Before that, there was a week-long conversation about confiscating “dust” UTXOs. Mark Karpelès, two episodes earlier, floated a fork to unfreeze coins tied to a court order so Mt. Gox creditors could be made whole. Amaury’s response, delivered with a grin: take it to BSV, they built that feature on purpose.

And beneath all of it sits the word he calls Bitcoin’s vague god: consensus. He dug up the old Theymos formulation from 2015, roughly “as long as one core dev disagrees, there is no consensus,” and diagnosed the trick: consensus talk is politics, and miners who fear wasting hash on a rejected block can be paralyzed by manufactured doubt, a self-fulfilling veto.

His own trolley-problem answer, if confiscation ever must happen: do not freeze the coins, recycle them as miner fees. Though he immediately notes the real problem is never writing the code; it is convincing thousands of anonymous strangers to run it.

As for eCash itself: the script system on XEC and BCH (unlike BTC’s, he says) is already rich enough for hash-based signatures, which are quantum resistant and deployable today with zero protocol changes. Not the ideal scheme, but a working fallback. The fancier lattice-based cryptography has a fatal practical flaw: public keys around two megabytes, meaning transactions roughly a hundred times larger. His doctrine: think about it now, have the plan ready, and do not shove half-baked post-quantum crypto down everyone’s throat before the threat is concrete.

Satoshi, the Multi-Chain Billionaire

A detail worth pausing on: Satoshi’s UTXOs exist, untouched, on BTC, BCH, BSV, and XEC alike. Every fork made him richer on paper. At roughly 1.1 million coins and the prices of the day, Satoshi sits somewhere around $75 billion, comfortably in the top ten wealthiest humans, assuming he is one human. “I don’t think money is his problem anymore,” Amaury deadpans.

This is also where we took a detour into AI-designed proteins (the same generative trick as image models, run in reverse) and the man who recently cooked up a custom treatment to save his dog from cancer. Amaury finds that frontier genuinely wild. So do I.

Wallets, Sponsors, and a Trade

The mid-show sponsor ritual had teeth this time. Cake Wallet, sponsor and prize provider, does not support XEC, which Amaury greeted with theatrical disappointment: “Oh, it sucks. Come on, Cake Wallet.” His engineering point stands, though: if you already support BCH, supporting XEC is nearly free, since the base protocol barely diverged. SideShift.ai does support XEC, which led us to Andreas Brekken, whom Amaury counts among his favorite Bitcoiners precisely because he never picked a tribe: he listed Lightning on SideShift earlier than almost anyone, ran into its limitations earlier than almost anyone, and sat in on Bitcoin Cash calls to see if that would work instead.

By the end of the show we had struck an explicit trade: I introduce him to Cake Wallet’s CEO about an XEC listing, he introduces me to Jason Dreyze, whose hash-based, quantum-resistant signature scheme for BCH recently shipped without any fork at all and, in Amaury’s telling, is probably not unrelated to BCH’s recent price strength.

Fractional Reserve Is a Throughput Technology

Buried in the middle of the episode is the tightest statement of Amaury’s economic worldview, and it is worth extracting. Why do banks exist in their current form? Because moving gold is slow, so early bankers issued claims on gold, and fractional reserve was born as a scaling solution for a low-throughput base layer.

It is not a conspiracy of mustache-twirling villains; it is the emergent answer to a velocity problem. Lightning, he notes, reinvented the exact same liquidity problems on purpose. And if you cap Bitcoin’s base layer at seven transactions per second, you have not escaped that history, you have condemned yourself to speedrun it: custodians, clearinghouses, fractional everything, rebuilt in a decade because the economics did not change.

Computers gave humanity, for the first time, the ability to raise monetary velocity at the base layer. Refusing that gift is a choice with a predictable ending, one he says he described at a conference in 2016 or 2017, in French, unfortunately for its fame: if Bitcoin does not scale, it will be completely absorbed by the existing financial system. A decade later: “It’s exactly what I said would happen.”

Drivechains: Right Problem, Slow Peg

On Paul Sztorc’s drivechains, Amaury is measured and genuinely interested. The problem BIP300 attacks is, in his view, one of the key unsolved problems in all of crypto: you cannot bake every feature into the consensus layer and get everyone to agree on everything, so blockchains need extension mechanisms.

Sidechains were explored exhaustively in Bitcoin’s early days and mostly abandoned because they are brutally hard. His specific criticism of the drivechain design is the withdrawal time: two consensus systems must acknowledge each other through mined confirmations, and bringing coins back to the main chain can take months, quite literally. Maybe that is fine if Bitcoin becomes the de facto standard and “good enough” wins, as it often does.

But eCash (XEC) has its own plan here, and it is the one place where Avalanche changes the game: because eCash has a second, much faster consensus layer on top of proof of work, the reverse peg can settle dramatically faster. The implementation work is substantial, but the architecture exists. Sztorc’s Ecash (ECX) launches its version of this vision tin August; Amaury’s critique is now on the record before it does.

What eCash (XEC) Actually Changed: Avalanche, Heartbeat, and the 51% Tax

Time for the technical core of the episode. eCash keeps the Bitcoin protocol essentially intact (same SHA256 mining, same 32MB block cap, same ten-minute target) and adds a second consensus layer running the Avalanche algorithm. The trade-offs are stated with unusual honesty. Avalanche reaches decisions in under two seconds, but it produces no auditable record: a freshly bootstrapped node cannot verify the history of past decisions the way proof of work’s accumulated chain allows. That is exactly why Nakamoto consensus stays put, and Avalanche is used where speed matters and where its weakness does not.

What does it buy? Post-consensus (which Amaury insists, correcting my chart, is not a hard fork): nodes continuously agree on whether a given block is in their preferred chain, and once finalized, a block will not be reorganized. Classic 51% attacks and a good chunk of selfish mining strategies simply stop working, because a late-arriving competing block is recognized as an attack rather than a maybe-I-missed-the-memo ambiguity.

Pre-consensus does the same for individual transactions: typical finality in one to two seconds. And heartbeat solves the minority-chain disease, where mercenary hashrate floods in when price pumps, jacks up difficulty, and vanishes, leaving hour-long gaps: blocks that arrive too fast get rejected, difficulty stops spiking, and block times end up as regular as Bitcoin’s despite eCash having a sliver of its hashrate.

The security economics produce the episode’s most cheerful boast. Attacking eCash now requires 51% of the hash rate and 51% of the staked coins. “If you want to do that, I’m fine with it. You’re going to have to make me absolutely rich. Buying that much of the supply will jack up the price quite a bit. Please do it.” People have already tried to 51% attack eCash, he adds. It did not work.

On staking rewards versus mining rewards, his answer is elegant and very minority-chain: there is nothing to balance, because the difficulty adjustment equilibrates profitability across SHA256 chains automatically. Pay stakers 99% and miners 1%, and mining simply gets a hundred times easier until returns match again. The market does the accounting.

Archaeology of the Block Size Wars

This is where the episode becomes a document. I pulled up a fan-made map of every Bitcoin fork and upgrade, and Amaury narrated and corrected his way through fifteen years of consensus history.

The pre-history: Bitcoin XT (Gavin Andresen’s BIP 101, the doubling schedule) and the era when supporting it got you DDoSed, including Slushpool, the first pool to signal for it, a fun detail given that Braiins, their modern incarnation, sponsors this very episode. Then Unlimited, then Classic, and the episode’s canonical example of big block self-sabotage: the joint testnet activation where XT and Unlimited failed to agree on the Sigops limit (a parameter so obscure that, as Amaury notes, most Bitcoin educators have never heard of it), allowing an attacker to split the effort into three testnets instead of two.

Two would have been a triumph. Three was a catastrophe for public trust. “You cannot split the chain in three when you plan to split it in two and expect people to run your software.”

Then the scoop, delivered casually, three and a half hours in: a private 2016 chat room, more than 50% of the global hash rate present, plus major industry players, plus the big block client developers, assembled to agree on the specific plan for activating larger blocks on Bitcoin itself. Everyone is converging nicely. Then the Bitcoin Unlimited contingent logs in, late, and lead developer Andrew Stone declares that coordinating the activation this way is collusion, and that BU’s “emergent consensus” mechanism must be used instead, a mechanism later shown unsound in an academic game theory paper (Amaury sent me the PDF mid-interview; the short version: the first mover gets punished, so nobody moves).

His verdict, quoted at length because it deserves it: “Imagine the degree of retardation. You’ve won the game. You have everyone who matters at the table. All you have to do is agree on the details of how you win. And you say: I don’t want to win the game. If you behave like that, you are never going to win any game. It’s guaranteed.” That meeting, he says, is the moment he concluded the existing big block leadership was beyond saving and started work on Plan B, which became Bitcoin Cash. He will not name the other attendees and no recording exists, so file this as one participant’s sworn testimony, but it rhymes with everything public.

The fork map proper: 2010’s OP_NOP hard fork, the value overflow bug that printed billions and Satoshi’s soft fork fix, the 1MB limit. 2012’s P2SH. 2013’s accidental Berkeley DB to LevelDB fork. 2015’s strict DER encoding (killing signature-encoding malleability) and CLTV, the last soft forks of a united community, whose timelock mechanics Amaury explains via Lightning’s punishment transactions. Then August 2017: BCH on the 1st, SegWit on the 24th.

Why BCH never took SegWit: because everything SegWit achieved could be done cleanly as a hard fork, and BCH did exactly that from day one, with a new transaction hashing algorithm that fixed malleability, fixed the quadratic hashing blowup (twice the transaction, four times the hashing work), and fixed the hardware wallet problem (offline devices previously could not verify input amounts, so a malicious companion app could trick a Ledger into signing away two bitcoin as “fee”). Three genuine fixes, one algorithm, no SegWit baggage. Meanwhile BTC has shipped exactly one soft fork since: Taproot, November 2021.

BCH’s launch difficulty algorithm “completely sucked,” and he owns why: to sell the fork, the DAA had to behave identically to Bitcoin’s in the majority-chain scenario, a constraint that made a good algorithm nearly impossible. Once BCH’s minority status was undeniable, the politics cleared and a sane algorithm (ASERT, in November 2020) replaced it. Bitcoin Clashic, the troll fork, gets one line and a smile. Monolith (May 2018) raised the cap to 32MB and restored the Satoshi-era opcodes in repaired, safe form.

Magnetic Anomaly (November 2018) added OP_CHECKDATASIG and, more importantly, CTOR, canonical transaction ordering, which he explains lovingly: ordering metadata for a block’s transactions grows much faster than the transaction list itself (one ordering for one transaction, six for three, and it explodes from there), so mandating a canonical order makes block propagation and parallel validation dramatically cheaper. Invisible to users, decisive for scaling. He also warns, as an aside that BCH merchants should tattoo somewhere: zero-conf on BCH is fine for coffee and not for cars. The ecosystem’s confidence in it exceeds its actual security; it works because nobody seriously attacks it.

The BSV Interlude: Speed Limits, Stolen Coins, and SQL Databases

Amaury does not pretend to be nice here, and I did not ask him to. The November 2018 split had one reason: “Craig is Satoshi, so whatever Craig says is gospel.” Craig Wright, he reminds the audience, is an Australian man who claims to be Satoshi Nakamoto and obviously is not, and not a competent engineer either, so the project he steered made changes it did not understand.

The Quasar block cap increases were, in his analogy, like abolishing speed limits between Paris and Bucharest: the road and the car determine how fast you can actually go, and BSV changed neither, which is why nodes faceplanted when the giant blocks arrived. The Genesis “restoration” removed P2SH and phased out CLTV/CSV as corruptions of Craig’s original vision, a framing he calls religious rather than engineering, and in the process created a cross-chain theft vector: any P2SH coin spent on BTC or BCH revealed the data needed to steal the mirror coins on BSV.

Then came the 2021 blacklist manager and the 2022 digital asset recovery confiscation feature, at which point his conclusion writes itself: if an entity can freeze and seize coins at will, you do not need a blockchain; you need a SQL database and a bank’s phone number. Why do people still believe Craig? “The con in con man stands for confidence.” He produces complicated-sounding explanations for everything; engineers hear nonsense, non-engineers hear brilliance. Amaury has been sued by that camp twice. Others have done better, he notes darkly, meaning worse.

Schnorr, CashTokens, and the Protocol-Branding Disease

The BCH upgrades that followed the BSV split get fairer treatment. Schnorr signatures (2019): simpler than ECDSA, batch-validatable (huge for block verification), enabling multisig that does not look like multisig, and only absent from Bitcoin’s early design because of a since-expired patent. Quantum resistance: identical, sadly; if one falls, both fall. Then the 2022 and 2023 crop: introspection opcodes, OP_MUL, bigger integers, CashTokens, P2SH32 (a bigger hash, because the original P2SH hash length is getting uncomfortably close to brute-forceable margins).

And here Amaury turns the knife on his former chain. CashTokens, in his assessment, is a dead end. The industry standard for tokens is ERC20, full stop; baking a non-standard token scheme into the consensus layer means eternal maintenance of something the market never asked for, at the slowest possible upgrade cadence, forever. “When I left, BCH was a top four coin. They focused on all this instead of making it the best cash in the world.”

The deeper pathology: when developers are not paid in money, they are paid in ego, and ego means tagging the consensus layer with your pet feature before you leave for the next chain. Bitcoin Unlimited’s people pushed their favorite ideas through, then departed to build Nexa. CashFusion, by contrast, is his Exhibit A for doing it right: a full privacy protocol deployed entirely outside consensus, which is why a second, better version could replace the first without any fork at all.

Hal Finney’s Banks, Read in Full, Endorsed in Full

The episode’s most disarming moment. I brought up the December 30, 2010 BitcoinTalk post in which Hal Finneypredicts Bitcoin-backed banks issuing digital cash redeemable for bitcoin, fractional or fully reserved, competing on policy à la George Selgin‘s free banking theory, with Bitcoin as the high-powered reserve money settling net transfers between institutions. Small blockers have weaponized this quote for fifteen years. Amaury found it, read it aloud in full, and delivered the verdict of the man who forked Bitcoin over the block size: “There is nothing in there that I would disagree with.”

His point is subtlety, not surrender: the quote does not make Finney a small blocker either. The man described secondary layers; he did not cap the base layer. Both camps, Amaury insists, put words in dead men’s mouths, which is exactly what he says about Satoshi’s big blocker status: the Garzik “don’t run that” 7MB patch proves Satoshi opposed big blocks then, in that way; the “temporary measure” comment about the 1MB limit proves he expected change, but he never specified what change; and his 2008 answer to James A. Donald, the very first reply the whitepaper ever received, invoking bandwidth growth and Moore’s law, rests on assumptions that stopped holding around a decade ago.

In the nineties, a three-year-old computer was a doorstop; today an M1 MacBook still hums along. Storage capacity ballooned while storage speed barely moved. “If Moore’s law no longer holds, would Satoshi have revised his position? You cannot know, because he is not saying anything anymore.” Honesty of this caliber about one’s own founding myth is rare in this industry, and it is the reason this episode ran four and a half hours without a wasted one.

Privacy: CashFusion, Auditability, and the Stupidest Part of the Whole Stack

CashFusion runs on XEC as well as BCH, one click in Electrum, and Amaury describes the cryptography plainly: participants pool coins, redistribute them in standardized amounts, nobody can cheat, and each round compounds your anonymity until you approach the privacy level of pre-confidential-transaction Monero, ring signatures without hidden amounts. He claims a chain analysis firm once examined CashFusion outputs and could not see through them. User error remains the weak link: consolidate your mixed UTXOs back into one and you have deanonymized yourself with your own hands. AI, in his reading, upgrades amateur analysts toward professional level, but it will not crack what professionals already cannot, because professional chain analysis was always algorithmic.

Would he hide amounts on eCash? He would like to, eventually, but the trade-off is real: hidden amounts kill supply auditability, the ability to verify nobody is secretly inflating. “Maybe Monero’s cryptography has been broken for years and someone is printing money every day, and we have no way to know.” The architecture he admires is Zcash’s: isolate the shielded pool from the auditable base, so each side keeps its guarantees. Exchange listings would get harder. He knows.

But his real passion here is a layer nobody discusses: the payment protocol itself is broken. Today, receiving money means publishing an address and having your wallet trawl a planetary bucket of everyone’s checks to find yours. It is, he says, an incredibly stupid design, bad for scaling and catastrophic for privacy, because every payment unnecessarily links inputs together in public.

A proper payment protocol, payer hands transaction directly to payee, combined with standardized, bill-like denominations, could deliver something like the practical privacy of physical cash (technically traceable, practically untraceable, acceptable to nearly everyone) without any consensus change at all. The chains we already have could do this. We simply never agreed on the protocol. File this under things to build.

The privacy coins get their report cards too: Monero’s UTXO set is effectively a TXO set that can never shrink (you cannot know which outputs are spent), a genuine scaling wall that even FCMP++ will not fully cure; Zcash pins its hopes on Tachyon’s client-side validation; Mimblewimble is elegant, amount-hiding, block-aggregating mathematics that nonetheless leaks everything to anyone who monitors the network live, and whose real appeal is aesthetic. It is a sports car for cryptographers; users want a Corolla that starts every morning.

The Nostr Question: Secret Agencies and the “Dev Tax”

Midway through hour three, I read him the hostile question of the day, from Nostr: alleged ties to secret agencies, and why was he so determined to impose a dev tax on BCH that largely benefited himself? He took both with visible amusement. No, he has no agency ties, while noting that this is exactly what an agent would say, and that being thought a secret agent is honestly kind of cool. As for the “tax”: a tax is collected by people with weapons who come to your house. Nobody was forced to run his software. The word choice, he suggests, reveals conceptual sloppiness.

The substance: ABC was the least-funded team in the BCH ecosystem while doing nearly all of the development, drowning in politicking and in other teams’ bad ideas, which consumed the scarce resources ABC did have. The proof of the model, he argues, is comparative output: since the split, his side delivered 51% attack protection, instant transactions, regular block times, and fork-free upgrades, the stated goals of Bitcoin Cash itself, “ten times more than BCH on those fronts, with a fraction of the ecosystem and resources.” Meanwhile BCH suffered the brain drain: BU gone, Classic gone, XT gone, the strong engineers gone. He believes, and says so plainly, that BCH would still be a top four coin on his roadmap.

He also offered the most humane explanation I have heard of why Gavin Andresen will never appear on your podcast: Amaury occupied the equivalent position at BCH and describes the incoming abuse as something you must multiply by ten from whatever you imagine, unpaid, without the formal mandate that lets a Vitalik pull rank. Wladimir van der Laan, by contrast, had a whole industry behind him. Multiply his complaints by ten and you approach Gavin’s reality. “I completely understand his position.”

Rogues’ Gallery: Roger Ver, Craig Wright, Epstein, Greg Maxwell, and Mark Karpelès

Roger Ver: “The best and the worst thing to happen to BCH at once.” A promoter of historic magnitude who also wanted everything yesterday, with no funding attached, bankrolled some of the ecosystem’s worst actors, and antagonized people Amaury then had to un-antagonize. They have not spoken in a long time.

Craig Wright was, for many who knew perfectly well he was not Satoshi, an ally of circumstance against ABC, until he forked and took his developers with him. The Epstein files renewed old knowledge rather than revealing new: the MIT Media Lab’s Epstein money and its employment of Gavin, Wladimir, and Cory Fields have been public for seven or eight years, and Amaury doubts it changed Bitcoin’s course, though he notes he personally refused dodgy money more than once, because once you take it, you do not know what you are in for.

Greg Maxwell still emails him, but about coordinating on Craig Wright lawsuits, not protocol design. And Mark Karpelès, freshly interviewed on this podcast, is building on Solana now; Amaury tells the longer version of the Mt. Gox story as a character study in corner-cutting, including, in his telling, the France episode that preceded Japan. His reading: negligence more than malice, but a pattern that was always going to collide with something, and it collided with the biggest thing possible.

Autopsies and Prophecies

SegWit2X gets its final autopsy: the deal paired an easy soft fork with a hard hard fork, activated the easy half first, and thereby destroyed the other side’s incentive to honor the second half. “Doomed to turn into a bait-and-switch.” His chicken-game metaphor from our first interview survives: one driver removed the steering wheel and floored it.

We agree to disagree on the counterfactual: I think BCH’s existence killed 2X; he notes Roger Ver only joined BCH after 2X collapsed. Either way, his verdict on the whole war is the one historians will quote: Ethereum won the block size wars. Frustrated builders and businesses did not wait; they left. He is “fairly confident Bitcoin would have big blocks today” if the pre-2017 big blockers had simply agreed on a plan.

My own prediction went on the record: two more Bitcoin hard forks within the year, one from the filtering faction (they already have the pool, the client, and the congregation) and one from frustrated builders with CTV and OP_CAT work and, in some cases, investors to answer to. Amaury’s counter is experience talking: Bitcoin is not the place to prove your tech. Deploy it elsewhere; if it is great, Bitcoin can adopt it later. Wanting it on Bitcoin specifically is ego, and he watched ego burn the big block movement to the ground.

Lightning gets the nuanced funeral: genuinely interesting technology, a mostly custodial reality (Phoenix and Electrum’s trampoline tricks acknowledged as the honorable exceptions), seamless custodial UX that hides the difference, and then the receipt, which I read aloud from page 55 of the Lightning paper itself: seven billion people, two channels each per year, 133 megabyte blocks. At that point in the evening, the document reading its own assumptions back to it felt like the episode in miniature.

Ten Years Out

Bitcoin: a captured reserve asset, price probably higher, every cypherpunk promise quietly retired, KYC on everything, absorbed into the system it was built to route around. BlackRock will never fork it, because Bitcoin’s failure to scale isthe product: custody is the business model. Bitcoin Cash: a genuinely good spot, not because of what is being built but because of a decade of integrations, with a BCH ETF likely on the way. eCash: hostage to the industry’s nihilist mood, waiting for the renaissance that always follows the death of the garbage, whether that is next year or in five.

On the rest of the market: the top ten probably persists, USDT remains eternally on the verge of exploding without exploding, Avalanche-the-coin can never be fully trustless (pure Avalanche consensus has no auditable bootstrap, unlike his hybrid), and Kaspa’s DAG runs into the speed of light itself: at ten blocks per second, miners cannot possibly know the previous blocks exist, so the branches buy miners rewards and buy users nothing. The real prize, he insists, is qualitative, not quantitative: get settlement under three to five seconds and you have built a different product, instant payment. Shaving block times never gets you there, which is why eCash keeps ten-minute blocks and puts the speed in the layer above.

Closing: Write Your Genesis Story

Before we hung up, Amaury plugged Cain’s writing contest at proofofwriting.com: a 100 million XEC prize for the best crypto genesis story, part of a book in progress. Follow the project at e.cash and @eCashOfficial, and Amaury himself as @deadalnix on X, GitHub, and everywhere else, including, apparently, my own mentions. My closing blessing stands as printed: flip BCH and prove everyone wrong. His answer: BCH has the ETFs and the penetration, so it will be difficult. “But you know, we’ll see.”

This episode is sponsored by Cake Wallet (1.75 million downloads and counting; the giveaway bank; XEC listing pending a certain introduction), SideShift.ai (supports XEC natively; swap whatever you won into whatever you actually want; not available in the US), Braiins Hashpower (rent SHA256 hashrate for lottery mining, for testing eCash’s attack resistance in a legally distinct manner, or theoretically for a 51% attack on BSV; don’t sue me Craig, I am just saying the service exists),

Bitcoin.com News (survivors of the block size war trenches, now with an orange logo, new management, and the invaluable Satoshi archive), and LayerTwo Labs (drivechains live on signet today: big blocks, Zcash-style privacy, assets, all waiting on BIP 300, and all about to get very real on mainnet). Special thanks to all the sponsors for making this freely available to everyone!

Listen to Amaury Séchet on YouTube, Spotify, Apple Podcasts & More!

Time stamps:

1. 00:01:19– Intro: the man who forked Bitcoin twice
2. 00:02:23– “You went French in the summer of 2017”
3. 00:03:22– XEC down, BCH sideways: network effects
4. 00:05:05– Memecoin nihilism & the casino culture
5. 00:06:55– Covenants, spam filters & quantum: the discourse spiral
6. 00:07:29– BIP 360 & freezing Satoshi’s coins
7. 00:09:42– Quantum: unsolved physics, oversold promises
8. 00:10:21– Google’s 2029 deadline & why they only threaten crypto
9. 00:13:09– Satoshi’s P2PK coins: the ultimate quantum bounty
10. 00:18:00– From block size taboo to confiscation proposals
11. 00:18:38– Karpelès’s Mt. Gox unfreezing pitch (try BSV)
12. 00:22:19– Consensus: the vague god everyone appeals to
13. 00:25:09– Quantum-resistant hash signatures on XEC & BCH today
14. 00:30:06– Satoshi’s untouched coins on BTC, BCH, BSV & XEC
15. 00:33:22– Cake Wallet doesn’t support XEC (yet); SideShift does
16. 00:36:11– A Bitcoin standard still needs crisis liquidity
17. 00:38:24– Fractional reserve exists because of base-layer throughput
18. 00:41:24– Drivechains: months-long withdrawals; eCash’s faster peg plan
19. 00:46:52– Avalanche explained: seconds-fast but no auditable record
20. 00:50:28– Heartbeat: steady blocks on a minority chain
21. 00:52:09– Attacking XEC needs 51% hash AND 51% of coins
22. 00:53:44– Slushpool signaled Bitcoin XT & got DDoS’d
23. 00:55:34– The three-way testnet split over the Sigops limit
24. 00:58:06– Emergent consensus & the game theory paper
25. 00:59:53– “Bitcoin is fiat currency written in C++” (sorry Saylor)
26. 01:02:38– Giveaway: Amaury’s number is 34; 33 wins 1 ZEC
27. 01:09:37– Was Satoshi a big blocker? Reading between the lines
28. 01:17:21– Hal Finney’s Bitcoin-backed banks: the full 2010 quote
29. 01:21:21– Szabo & Ou’s radio Bitcoin: the 300KB block idea
30. 01:27:02– Calin Culianu: “he hates me” (but Fulcrum is good)
31. 01:30:29– Why every fork starts with an Electrum port
32. 01:33:40– 2010: the value overflow bug & the 1MB limit
33. 01:35:11– Strict DER encoding & malleability
34. 01:37:06– CLTV/CSV & how Lightning punishment works
35. 01:41:38– Why BCH never adopted SegWit: the sighash fix
36. 01:46:08– BCH’s first DAA “completely sucked” (by political necessity)
37. 01:48:52– Monolith: 32MB & resurrected Satoshi opcodes
38. 01:50:08– OP_CHECKDATASIG & CTOR: the underrated scaling upgrade
39. 01:56:29– Zero-conf on BCH: fine for coffee, not for cars
40. 01:56:58– BSV: the “Craig is Satoshi” fork
41. 01:58:25– Gigabyte blocks & the speed limit analogy
42. 02:00:22– BSV Genesis: unlimited blocks, P2SH removed, coins stolen
43. 02:04:35– Blacklists & confiscation: just use a SQL database
44. 02:07:21– Why people believe Craig: con man means confidence
45. 02:09:15– Kaminsky’s 2011 Visa math: 1GB blocks, 20 cores
46. 02:14:43– Schnorr on BCH: the patent finally expired
47. 02:19:54– Why the name eCash: Chaum, ecash.pdf & the e.cash domain
48. 02:26:22– BCH keeps shipping: OP_MUL, CashTokens, P2SH32
49. 02:27:47– The irony: BCH now has better L2 rails than BTC
50. 02:29:35– The 8% coinbase rule: the real reason for the split
51. 02:31:02– “If you’re not paid by money, you’re paid by ego”
52. 02:33:52– Regrets? “It’s not clear there was another path”
53. 02:36:21– The Coinbase listing night: BCH at #2
54. 02:40:22– SegWit2X was doomed to be a bait-and-switch
55. 02:42:26– Avalanche post-consensus: finality without forks
56. 02:45:49– Fork-free upgrades via Avalanche
57. 02:48:57– CashTokens critique: ERC20 won & consensus is forever
58. 02:54:12– CashFusion: privacy done right, outside consensus
59. 03:02:32– Hiding amounts means giving up supply auditability
60. 03:04:06– The payment protocol is broken: checks in a giant bucket
61. 03:07:46– Monero’s bloat: a UTXO set that never shrinks
62. 03:09:52– Mimblewimble: elegant, but watchable live
63. 03:13:09– The 100M XEC writing contest at proofofwriting.com
64. 03:15:24– Pre-consensus: 1-2 second finality
65. 03:19:58– Ordinals on XEC? Let Avalanche nodes vote
66. 03:21:34– Why nobody forks Bitcoin anymore
67. 03:23:24– BlackRock will never fork Bitcoin: custody is the product
68. 03:30:37– The Nostr question: secret agencies & the “dev tax”
69. 03:37:04– Gavin’s burden: multiply the abuse by ten
70. 03:41:15– The secret 2016 meeting: 50% of hashrate at the table
71. 03:43:16– Andrew Stone: “This is collusion” & the game thrown away
72. 03:48:58– Ethereum: the real winner of the block size wars
73. 03:51:40– Would SegWit2X have happened without BCH?
74. 03:56:21– Vlad’s prediction: two more BTC hard forks this year
75. 04:00:04– Lightning: brilliant tech, mostly custodial in practice
76. 04:01:29– Page 55: the 133MB blocks requirement, read aloud
77. 04:06:37– Craig Wright as an ally of circumstance
78. 04:10:00– Epstein, MIT Media Lab & Blockstream
79. 04:13:32– Karpelès builds on Solana; a history of cutting corners
80. 04:16:53– Roger Ver: the best and worst thing for BCH
81. 04:18:15– 10-year predictions: BTC absorbed, BCH ETF, XEC renaissance
82. 04:24:32– Kaspa’s DAG vs the speed of light
83. 04:31:18– Where to follow Amaury & eCash