S17 E22: Jeremy Rubin on CTV, Char & Relentless Building

Four years. That is how long I have been chasing Jeremy Rubin for an interview, and it took the Bitcoin 2026 conference in Las Vegas, a borrowed corner of the LayerTwo Labs booth (the one with the “evolve or die” banner staring into the camera), and a fully charged battery to finally make it happen.

If you do not know Jeremy, you have been living under a rock the size of a block template: he has been in Bitcoin since 2011, he is the author of BIP 119 (OP_CTV, Check Template Verify), co-author of BIP 348 (CSFS, Check Sig From Stack), founder of the MIT Digital Currency Initiative, and more recently the CEO of Char Network, a decentralized shared sequencer for Bitcoin layer 2s. He is also, depending on which corner of Bitcoin Twitter you ask, either one of the most important protocol thinkers of his generation or a difficult spook from MIT. We get into all of it.

Quick housekeeping: this was recorded live on the conference floor, so expect background noise, a very loud astronomy panel at the worst possible moment, Curtis from the Self Custody podcast walking by exactly when I needed him, and at least one Pokemon biology lesson. Cake Wallet sent 1 ZEC as a thank-you gift, which I had to send off camera because Jeremy has a hard rule we will get to. Nobody paid to be here and nobody gets editorial approval.

The Scientist Who Wrote CTV

Jeremy frames himself first as a scientist and philosopher of how Bitcoin should work as a tool for human freedom, and only then as an engineer. That framing matters, because it explains his relationship with his own creation: “I don’t really particularly care that much about CTV. What I care about is that Bitcoin can be used in the context where people need to use it.” His bet is that any well-meaning engineer looking at Bitcoin, today or five years ago, would independently arrive at something like CTV. Seven years after the technique was first described, under its former names COSHV and Secure the Bag, the design has barely moved. Not because it is stagnant, but because it was right.

The simplest explanation I ever got for CTV: it scales Bitcoin in time rather than in space. No bigger blocks, just the ability to program transactions so they land when fees are optimal. The flagship example is congestion control: if every Coinbase user wanted to withdraw at once, it would genuinely take months to process, and Coinbase would be honest saying so. With CTV, all of those payments collapse into basically a single transaction, and every user gets a verifiable certified check to claim their money from.

Then there are vaults, and Jeremy’s favorite variant, the vacation wallet: your cold setup at home can trigger a withdrawal to your hot wallet that takes a month to become spendable, and if someone breaks in while you are sipping margaritas, you can cancel and slam everything into deep cold storage. Theft prevention through time control. If that is not a vault, the word has lost all meaning.

The 2022 Activation That Wasn’t

Regular listeners know I have probably interviewed the largest collection of frustrated Bitcoin developers in podcasting, people waiting on this one upgrade so they can start building. Moonsettler, who nowadays goes by Doomsettler, spent most of his ten-hour episode complaining that we never got CTV. And the tragic part is how close it was: speedy trial, real momentum, and then the Telegram group happened. My reading of the archive was that Adam Back joining the chat and posting FUD for two weeks demoralized everyone into giving up.

Jeremy corrects the record on one point: his sabbatical was burnout, not CTV, and he kept working for nine months after the activation effort wound down. But his account of the social dynamics matches what I read in the logs, and it is damning in a quiet way. A paternalistic “I’m the owner of the protocol, jump and I’ll tell you if it’s high enough” posture. A total refusal to define “technical consensus” beyond “you’ll know when you have it.”

His utxos.org signals page had around 150 developers, business people and protocol researchers calling CTV a slam dunk, more public support than even Taproot had, and the response was essentially: that is not the kind of consensus we meant. Then came the FUD campaign, the absurdist claim that covenants are a government censorship opcode, amplified by people who really should have known better. Jeremy singles out a two-hour podcast by Andreas Antonopoulos warning that covenants would let the government own Bitcoin, while admitting he had not looked at CTV at all.

“You could have spent one hour reading the BIP,” Jeremy says, still visibly annoyed years later. The final twist: after CTV was defeated, the victors promised to “do it the right way.” Years passed, people went looking for the right way, found nothing, and quietly concluded the original criticism was never technical. That burned the technical credibility of the critics more than it burned Jeremy.

Blockstream: The Company and the Brain of the Organization

My personal theory, formed while reading those logs, was that Blockstream stalled CTV because they wanted to ship their own version later, dressed up as the Great Script Restoration. Jeremy is more charitable and more precise. Blockstream’s people gave CTV its most thorough review: Russell O’Connor (it was his feedback that drove the iterations from COSHV to Secure the Bag to CTV) and Greg Maxwell both contributed substantially. The problem, in his telling, is cultural: a belief in inherent technical supremacy, the assumption that anyone worth their salt will eventually work for Blockstream.

They tried to hire Jeremy multiple times, starting in college, and he kept refusing, not out of dislike, but because he had just helped set up the MIT DCI as an independent home for Core developers, and going to work for the company that already held so much of the technical talent would have defeated the point. His read: the organization never quite metabolized the bright kid who would not sit at their table. “Organizations and the people in them can be two separate things,” he says. He is still friends with many of the individuals.

And the replacement effort? Rusty Russell announced he finally understood CTV, would fix Jeremy’s mistakes, and shipped a “simplified CTV” that was, in Jeremy’s telling, completely broken. Then broken again after the first fix, which Jeremy demonstrated with a code sample. Then came the Great Script Restoration, a one-year effort announced with much fanfare at a bitcoin++ event, which fizzled out around three years ago. Meanwhile, CTV just sat there, still correct, still unactivated. Jeremy’s management critique is surgical: if your engineer publicly subverts momentum behind an upgrade by promising an alternative, it is now the company’s job to follow through, because the company’s reputation is bound to the promise. Nobody at Blockstream seems to hold those promises.

So What Do Jeremy and Adam Back Have in Common?

Here is the part you have been waiting for, and the part I teased at the top as the elephant in the room. Mid-rant about Blockstream, I asked Jeremy what he and Adam Back have in common these days. His on-camera answer was wholesome: they are both at the Bitcoin 2026 conference, Adam literally visible from where we sat. “He’s over there at that little island over there,” Jeremy said, pointing at the booth structure, and I need you to know he chose the word island completely innocently, which made it the single funniest unscripted moment of the episode.

The real answer, which listeners of S17 E7 with Colin Harper already know: both of their names appear in the recently released Epstein emails. To be crystal clear, because the internet will not be: appearing in those files is not an accusation of anything. Adam “Andy” Back was in Epstein’s orbit during the Blockstream fundraising era. Jeremy was an MIT student whose department was financed by Epstein money, and whose crime was emailing the financier of his department. He never went to any island, never took a ride on the Lolita Express, and his involvement is roughly as scandalous as a grad student thanking a donor.

The joke writes itself and the guilt-by-association writes nothing at all, but since half the conference was gossiping about the files anyway, I figured we might as well get the only two guys on the record about it on the record. For the record, I initially thought the files said “Andy Back,” which tells you everything about how carefully people read primary sources before forming opinions.

Institutionalized: Paper Bitcoin Vegas

Walking the conference floor did not lift our spirits. Almost no booths accept Bitcoin; the bar takes your credit card. Cash App’s booth was giving away $21, but only if you install their app and hold a US bank account, which as a European I took as targeted discrimination. The night before, Michael Saylor was on stage telling people to stake their stablecoins on Solana or Ethereum for yield.

Jeremy’s diagnosis is structural: when something is easy to deploy on-chain, you get meaningful decentralized infrastructure; when it is hard, everything routes through centralized platforms. His example is Uniswap on Ethereum: a fully decentralized, no-red-button token swap engine that people actually use. Meanwhile Blockstream tried to build something similar on Liquid and it is, in his words, super complicated.

The cypherpunk movement, he notes, is not really at these booths. Commerce is happening, but it is not settling on Bitcoin, and he wants everything to settle on Bitcoin. That, he reminds us, is what Bitcoin maximalism used to mean: everything outside Bitcoin is an experiment whose results get brought home. Now it just means being toxic.

He also shared a gem from his bitcoin++ panel with Peter Todd (footage courtesy of Lisa and Nifty, please release it): after dismissing every covenant proposal as “not how you would design it,” Peter declared that what you actually want is a programming environment that lets you encode state machines with UTXOs. Jeremy looked at him and said: “What do you think covenants are? That is literally what we have been building.” People care more about saying something bad about Ethereum than something good about progress inside Bitcoin, and that inversion is rotting the culture.

The Small Blocker Paradox and the Bitcoin Cash Confession

I said the thing that would get me punched by half the conference: Bitcoin Cash did a better job scaling Bitcoin, not through big blocks but through covenants, a token protocol, and the restored Satoshi opcodes, including OP_CAT. Jeremy, in heavy irony quotes the “author” of CSFS, confirmed that BIP 348 is largely an adaptation of Bitcoin Cash’s checkdatasig and the Elements opcodes, work that was done years ago and has sat unshipped ever since.

His summary of the situation deserves to be quoted in full: “We are the small blockers. We are the people who said no, we don’t need more data, we will figure out how to use it better. And now we’re not doing anything. So just make the blocks bigger. It’s so much worse to have no direction toward scaling than to take some additional data.”

Bizarro world indeed, and yes, a Raspberry Pi can still sync a BCH node, which suggests the block size war was more personal than technical all along. Fun fact he appreciated: Wasabi’s WabiSabi protocol took inspiration from Bitcoin Cash’s CashFusion. The big blockers were not the technical illiterates of legend.

So why don’t the covenant builders just build on Bitcoin Cash and prove the tech? Jeremy’s answers: evidence from BCH reads as anecdotal to Bitcoiners, the cultural norm routes serious builders to Liquid (a ghost town where blocks rarely carry more than two transactions, but at least nobody judges you), and there is a genuine perception that BCH exists for the enrichment of a specific set of people.

My counter-spin: building on Bitcoin today mostly enriches the establishment we used to pretend to hate. Jeremy’s deeper point is the hopeful one: “If I build on Bitcoin, maybe I can save Bitcoin.” A C program can do anything, but it does not run on Bitcoin, and people want to write programs that run on Bitcoin. We do these things not because they are easy, etcetera.

Ecash, Jealousy Game Theory and a Network Made of People

On Paul Sztorc’s Ecash hard fork, and I did warn him we were sitting in Paul’s booth, Jeremy was measured: Paul has every right to be as frustrated as he is, he is clearly well-meaning (there were many easier ways for him to make money over the years), but Jeremy does not believe the fork is good for the community in the way Paul thinks it is. He wishes him luck and wishes to be shown wrong. My framing was game theory: when you cannot get the girl, you date her best friend and make her jealous, and maybe Ecash triggers the dev exodus that forces Bitcoin to adapt.

Jeremy pushed back with the most important sociological observation of the episode: this is a social community, a network made of people, not machines, and you do not popularize an idea by doing the most unpopular possible version of it. His analogy: a gun rights activist does not win hearts by bragging about legally shooting twenty people, even if the Supreme Court would side with him. Even a perfectly sound technical evaluation body does not exist to absorb Ecash’s evidence neutrally, so the fork probably makes the case harder, not easier. The one thing that might work in Paul’s favor: launching something people have to care about, even if only to dump it, forces them through the technology itself.

On the SHAD (confiscating half of Satoshi’s airdropped coins to fund development), Jeremy calls it a bad look and offers the obvious improvement: take the coins only if they have not moved in ten years, give Satoshi a chance. His slippery slope is the billionaire’s tax argument applied to Bitcoin: a confiscation that applies only to Satoshi today applies to Hal Finney tomorrow and to you eventually. He is equally unsupportive of BIP 361, the quantum-motivated freeze, dismissing the whole quantum panic as conjecture about unlikely bad events and countering with the most Jeremy line possible: “I’d rather ship something like CTV that we know does something good today.” Smooth. I told him I should learn sponsor plugs from him.

Would he deploy his CTV work on Ecash to prove Bitcoiners wrong? No. If that were the goal, he would have done it on Bitcoin Cash already. The point was never to show it somewhere. It is to show it on Bitcoin.

OP_CAT: A Loaded Gun You Pick Up by the Trigger

Since people keep pitting OP_CAT against CTV (Andrew Poelstra once mused on this show about OP_CAT next year and CTV maybe in four), I asked for his real take. He is not opposed to OP_CAT, but the version on the table bothers him: the 520-byte stack limit cripples the covenants you would actually want, capping a covenant transaction at around thirteen outputs because of how the signature digest fields hash.

His deeper critique is a culpability argument I have not heard anywhere else: giving developers OP_CAT and calling it covenants is handing them a loaded gun and telling them to pick it up by the trigger. It is nearly impossible to write a correct OP_CAT contract, and very easy to write an incorrect one. Core’s posture of “we don’t want responsibility for a bug in Bitcoin Core” just pushes the bugs to the user level, and your average user does not care whether a consensus bug or a wallet bug took their money. Either way, the money is gone. His preference: ship specific, easy-to-use primitives. Like, you know. CTV.

Why Jeremy Will Not Install Your App at a Conference

When I did the Cake Wallet giveaway bit, Jeremy declined on grounds that should become industry standard: he does not install apps at conferences, full stop. Not because Cake is suspect (he likes the team and promised to install it at home), but because in the age of Mythos-grade LLMs and professional social engineers, conference-floor installs are how people lose everything. His cautionary tale: Gavin Andresen was socially engineered into believing he had validated a signature and still does not fully know how it happened.

If a mentalist can make you pick the card they want, a booth QR code can make you sideload the wrong APK. So the 1 ZEC thank-you from Cake was sent off camera. And no, this is not his first Zcash; there is a lost wallet from the Sprout era, a tragic boating accident like everyone else’s.

Char: Bitcoin-Native Sequencing, Named After Burning

Which brings us to what Jeremy is actually building now. Char (yes, like Charmander, who evolves into Charmeleon and then Charizard, and no, Jeremy does not keep up with Mega evolutions) is named after burning: lock up bitcoin, sign messages, and if you ever sign two conflicting messages, your key is revealed and a miner sweeps your bond.From that primitive, an oracle that can only attest one version per query, you build a proof-of-stake network, and from that you get pre-commitments: layer 2s that eventually post their data to Bitcoin can enjoy better-than-one-block latency, compress what they post, and skip posting every block.

Centralized sequencers can do this too, but they get special privileges, including transaction censorship, and Jeremy’s view is that those privileges should not belong to anyone. Hence a decentralized shared sequencer, citing Johnston’s law (everything that can be decentralized will be decentralized) even while noting he is not a fan of David A. Johnston’s broader work. The repo is public at char.network, open for review, iterating in the open.

Why proof of stake in a culture that treats it as a perpetual motion heresy? “Everything at the end of the day seems to be based on memes,” he says. Proof of stake is just a technology. Char does not turn Bitcoin into a nothing-at-stake system; it is a careful, time-bounded adjudication layer anchored back to Bitcoin.

His closing argument is elegant: if you refuse to build on the block you just mined, you lose bitcoin. If you double-sign in Char, you lose bitcoin. The punishments are denominated in the same asset, so the economic security guarantee is the same. When he hedged with “if we do it correctly,” I pushed, and he corrected himself with a grin: “We have done it correctly. But I am a scientist. Show us we are wrong and we will evaluate the claim.”

The Spook Question and the Protagonist Speech

I could not let him leave without addressing the rumors: people had warned me he is difficult in person, or possibly a spook because he worked at MIT. His answer to the spook accusation is the best possible one: “That’s not what you said when you were talking to your wife last night.” (He was listening in. That is the joke.) More seriously: there is nothing he can say that would dissuade a believer, so the only move is to keep working on freedom technology.

On being difficult: he says what he thinks, dissent included, and Bitcoin needs to relearn that work disagreements are not personal disagreements. Considering the character assassination he has absorbed over the years, the lack of bitterness is remarkable. He got into Bitcoin in 2011 as an audit-the-Fed Ron Paul Republican going to rallies, because monetary policy was the issue he thought could destroy the society he loves. If TXHASH activates instead of CTV, no skin off his back. It was never about ego.

His closing message, which I will not even try to top: find where you have agency, find where your decisions can have the impact you want, and do what a protagonist would do. Do not live in the shadows, unless that is where your protagonist lives. A gentleman and a scholar, even if I scrambled the compliment on air.

This episode is sponsored by LayerTwo Labs (our hosts at the booth; evolve or die, test the drivechains today, Ecash launches at block 964000 in August, details at ecash.com), Orange Rock (the no-KYC trading platform from the Bitcoin.com team, with leverage for the gamblers, which I do not recommend), Braiins Hashpower(rent hash rate for lottery-ticket mining, for the new altcoin Paul is launching in August, or, theoretically, to 51% attack BSV; I am not saying you should, Craig, I am just saying you can), Sideshift.ai (when someone offers to pay you in stablecoins, swap that into BTC; not available in the US), and Cake Wallet (the gateway drug to Monero with excellent Zcash, and the sponsor of Jeremy’s 1 ZEC gift, delivered off camera like responsible adults). Special thanks to all the sponsors for making this freely available to everyone!

Listen to Jeremy Rubin on YouTube, Spotify, Apple Podcasts & more!

Watch on YouTube

Time stamps:

00:00:00 – Intro music & heat check
00:00:48 – Podcast intro & the sponsor gauntlet
00:02:10 – Saylor’s stablecoin staking vs Sideshift
00:02:37 – Braiins: rent hash rate for Paul’s August altcoin
00:03:02 – Four years of chasing Jeremy
00:03:24 – Fifteen years in Bitcoin
00:03:41 – The elephant in the room & CTV, formerly Secure the Bag
00:04:28 – CTV + CSFS before the Knots takeover
00:04:44 – Jeremy the scientist: freedom is the goal, not CTV
00:06:19 – Scaling in time instead of space
00:06:40 – Congestion control: the Coinbase bank run example
00:07:57 – The vacation wallet: vaults without the buzzword
00:09:35 – Moonsettler’s 10-hour CTV grievance
00:10:44 – 2022: speedy trial & the Telegram group
00:11:17 – The sabbatical was burnout, not CTV
00:11:25 – “Technical consensus”: you’ll know when you have it
00:13:11 – 150 signers on utxos.org meant nothing
00:13:45 – The government censorship opcode FUD
00:14:42 – A two-hour podcast without reading the BIP
00:16:00 – “Where’s the right way?”
00:16:33 – Vlad’s theory: Blockstream wanted their own CTV
00:16:47 – What Jeremy & Adam Back have in common
00:17:05 – “He’s at that little island over there”
00:17:45 – Blockstream: real contributions, supremacy complex
00:19:23 – Russell O’Connor’s thorough review
00:20:06 – The hiring attempts & why MIT DCI stayed independent
00:21:44 – Rusty Russell’s broken simplified CTV
00:22:54 – Great Script Restoration: three years of fizzle
00:24:20 – Institutionalized: credit cards at a Bitcoin conference
00:24:51 – The $21 Cash App booth needs a US bank account
00:25:40 – Saylor, Solana staking & paper Bitcoin
00:26:08 – Hard base layer, centralized everything
00:26:51 – The Uniswap example
00:28:10 – The cypherpunks are not at these booths
00:28:40 – What Bitcoin maximalism used to mean
00:29:52 – The Peter Todd panel: “what do you think covenants are?”
00:31:19 – The intellectual backflips of BitVM
00:31:52 – Bitcoin Cash did a better job scaling (fight me)
00:33:55 – CSFS adapted from BCH’s checkdatasig
00:35:39 – The small blocker paradox: “we’re not doing anything”
00:36:12 – Just make the blocks bigger: bizarro world
00:36:42 – A Raspberry Pi can sync a BCH node
00:37:34 – Why won’t covenant builders touch Bitcoin Cash?
00:38:41 – Liquid: the acceptable ghost town
00:39:22 – Enrich the libertarians or the establishment?
00:39:47 – “If I build on Bitcoin, maybe I can save Bitcoin”
00:40:12 – Programs have to run on Bitcoin (the JFK doctrine)
00:41:04 – Monero, Zcash & stablecoins eating our lunch
00:42:12 – The window of opportunity is narrowing
00:43:08 – Jeremy on Paul’s Ecash fork (“not in his booth, man”)
00:44:35 – The make-her-jealous game theory
00:45:31 – It’s a network of people, not machines
00:46:47 – The gun rights activist analogy
00:49:06 – Taking Satoshi’s coins is a bad look
00:49:56 – Satoshi today, Hal Finney tomorrow
00:50:39 – BIP 361 & the quantum freeze debate
00:52:19 – Quantum conjecture vs shipping CTV today
00:52:54 – “Let people program Bitcoin”
00:53:54 – Would Jeremy deploy CTV on Ecash? No.
00:55:22 – Forks as free-market experiments with real incentives
00:55:54 – The 21M cap is not in the whitepaper
00:56:14 – BIP 42: a relic of a prior era
00:57:06 – Bitcoin invites engineers to compromise their values
00:58:10 – OP_CAT & the 520-byte stack limit
00:59:55 – A loaded gun you pick up by the trigger
01:00:54 – Users don’t care whose bug stole their money
01:01:56 – The Cake Wallet question & Jeremy’s hard rule
01:02:34 – Never install apps at conferences
01:03:30 – Sideloading, mentalists & Gavin’s signature
01:04:24 – He still wants the Zcash
01:04:30 – Zcash: what Bitcoin should have become
01:06:01 – What excites Jeremy: Char & the L2 breakthrough streak
01:06:46 – Robin’s quantum-resistant hash signatures
01:08:16 – Post-burnout: only work that feels good
01:09:30 – “That’s not what you said to your wife last night”
01:10:38 – Work disagreements are not personal
01:12:20 – What is Char (and why it sounds like Charmander)
01:13:07 – Lock BTC, sign once, or the miner sweeps
01:13:39 – Pre-commitments: better-than-one-block latency
01:14:17 – Why the sequencer must be decentralized
01:15:21 – Proof of stake is just technology
01:17:20 – Cheat in Char, lose Bitcoin
01:18:12 – The scientist’s caveat: “we have done it correctly”
01:18:39 – char.network & why he’s not bitter about CTV
01:19:29 – Ron Paul rallies & auditing the Fed
01:20:24 – Bitcoin has the most potential: don’t squander it
01:21:13 – Agency: live like a protagonist
01:22:25 – Sending the ZEC off camera
01:22:37 – Outro & the 51% attack disclaimer (don’t sue me, Craig)