S15 E66: Jameson Lopp on Michael Saylor, Bitcoin Ossification & Paranoid Crypto Anarchism

I sat down with Jameson Lopp at NiceHash X in Maribor, Slovenia, and opened bluntly. We like getting cancelled on Twitter, I told him, and we do not like Michael Saylor. Lopp would not sign the manifesto. He would agree with Saylor on some things and disagree on others, and that was fine, as long as Saylor does not jeopardize the future of Bitcoin. In his telling, Saylor is playing a political game and pushing a narrative about what the project is that neither of us shares. He pointed at Saylor's own framing: "I think in one of Saylor's recent interviews he definitely differentiated himself from the paranoid crypto anarchists."

That was the October 2024 conversation with NZ Herald's Madison Reidy, the one where Saylor waved away seizure fears around regulated custodians as the hang-up of paranoid crypto anarchists. I asked Lopp if he was one of them, and if Casa, the self-custody company he co-founded, was built for that tribe.

"I would call myself a paranoid crypto anarchist. Casa is somewhere in the middle."

Casa does not require KYC, and people can work with the company anonymously or disclose more information for authentication. "We're flexible as a company," he said. "Casa is really here to help people help themselves, and especially the non-technical people, the folks who want better security but maybe they just don't have the time and effort to climb the mountain that is required to understand all of the different decisions and tradeoffs around self-custody."

This was his second time on the show. In Season 14 Episode 9 we talked about Casa, self-custody, and Bitcoin maximalism. A little over a year later the job had not changed. The weather around Bitcoin had: exchange-traded funds, a corporate treasury class, and a choir that wants the base layer to freeze in place. Lopp is one of the faces people paste onto maximalism, the guy in the Make Bitcoin Great Again hat. He does not talk like the people who think that picture is a theology.

133 Megabyte Blocks, If We Have Not Ossified Already

I reminded him of a Bitcoin Magazine piece from about eight years earlier in which he had talked about Bitcoin needing something like 128 megabyte blocks. The number, he said, is not a personal fantasy. "Usually that's the number I pull from the Lightning Network white paper, right? 133 megabyte blocks required for global adoption of the Lightning Network if we assume, I think, two channel opens and closes per year per person." Those assumptions sit in the Poon-Dryja paper, and they are not laws of physics. Shared UTXOs and multi-party channels could cut that demand by orders of magnitude. "The future is not set in stone, or at least it might not be if we haven't ossified it already."

Should Bitcoin just stop? Send people to Ethereum for contracts, Litecoin for low fees, Monero for privacy? Is that a surrender of what maximalism used to mean, which was not "this is already perfect"?

He found the ossification camp stuck in a contradiction. "I would say probably most of the hardcore ossification proponents would not say go build on Ethereum or use other networks, because they tend to be purists and think that people should only be using Bitcoin. So that's kind of stuck between a rock and a hard place." If Bitcoin cannot keep adding functionality and the ability to create more layers, "then I think that you should expect that people are going to go use other networks, other protocols that fit those use cases."

Mental Gymnastics on the Second Layer

The popular escape hatch does not impress him. "This is a common trope in the space of, oh, we don't need to change the base layer because we can just do everything on other layers, and that's technically true with enough mental gymnastics."

Those gymnastics keep failing to spit out a permissionless second layer. The things that ship most often are fully custodial (he would put a lot of e-cash mints in that bucket, with real pros and cons) or federated pegs, or else "really complicated and convoluted game theory" that tries to make a peg more trustworthy.

Lightning is already a complicated specimen of that game theory. So is Botanix and its Spiderchain, "this series of multisigs where people are pegging in and pegging out the funds," a rolling set of wallets controlled by different groups of stakers, designed so there is no single point of failure. The same week, a covenants paper had landed that tries to conjure new behavior out of hash collisions. He named Ethan Heilman and Andrew Poelstra. The idea is fascinating as an application of tools Bitcoin developers already have. He does not think it will get used anytime soon. "It's incredibly expensive and impractical. I think the current iteration of it basically requires tens of millions of dollars worth of electricity to be expended in order to find a colliding hash that can allow you to recreate this covenant functionality."

Developers are burning time on designs that probably will not be adopted, "mainly because people or developers are so pessimistic about even being able to soft fork in a new opcode that could add powerful functionality."

Seven Years of Lightning, and Most of the Zaps Are Custodial

Lightning went live on mainnet in 2017, so we were about seven years in. I have not been kind to it on this show. Adoption looked underwhelming after 2021, and I watched the sovereign dream slide toward LSPs, Wallet of Satoshi, and Cashu. I used to run a node over Tor and I think I lost more money than I made.

Lopp did not pretend there would be one ultimate scaling solution. Casa was an early adopter and "actually had one of the first Lightning node in a box." They only did it for a few years. "The support costs were through the roof." He had been one of the hobbyists with grandiose hopes, carefully maintaining a node and trying to optimize liquidity so other people would route through him.

"Almost nobody who's operating a node on the Lightning Network is actually profitable. It's probably like 1 to 5% of node operators that are collecting enough fees to offset the things like force channel closures that can be very expensive."

Nostr helped, because people actually zap each other. The statistic he cited was less romantic. "I think 80% of zaps are happening through custodial wallets." Liquidity management is still too hard, so most people outsource the node. At the Lightning Conference in Berlin in 2018 or 2019 he argued that exchanges needed to get on Lightning. Years later he could count only two or three. If they used unified payment codes for on-chain and off-chain deposits, he bets on-chain demand would drop, because "about 50% of on-chain transactions are just people sending from one exchange to the other." Until better tooling exists, LSPs remain "a hacky workaround."

Drivechains, Restored Opcodes, and a Culture That Got Comfortable

I brought up fiatjaf, who had told me to burn Lightning and adopt Drivechains, a tip-sized sidechain you could prune every six months. The show is sponsored by LayerTwo Labs, so I asked.

"I'm not against Drivechains. I've never been convinced by some of the fears that people have said that Drivechains could totally upset the balance and game theory of mining on the main chain." He also does not expect BIP 300 or BIP 301 to get consensus as written. A roundabout path might appear if Bitcoin restores some of the opcodes that were taken away, leading to "layer twos that are real layer twos and not just like multisigs."

I used the opening to talk about Paul's CUSF idea, a Core Untouched Soft Fork aimed at miners, and I listed what a sidechain world could hold, including an Ethereum side chain or a Zcash side chain. That was my pitch, not his. Lopp did not follow me into Zcash.

The deeper problem, I said, is cultural: we went to the institutions and everyone stayed happy while the number went up, even though the $77,000 all-time high that day looked less impressive if you adjusted 2021's $69,000 for inflation.

Lopp called the anti-change choir a very vocal minority, running on "if it ain't broke, don't touch it." His pushback was practical. "If you can't adapt the protocol to the world as it's changing around it, there are a number of things that can go wrong." Bitcoin was at a nominal all-time high. "And yet if you look at pretty much any of the other metrics of the network, it's very much a bear market. Nobody is transacting. The fee rate I think is like two satoshi per vbyte right now. It's basically at the floor."

Institutions and ETFs opened new liquidity, but those buyers are not doing self-custody, so they are not making on-chain transactions. "We're just seeing billions and billions of dollars being funneled into the hands of a couple of custodians." His answer was the same one he has been repeating for a decade: promote self-custody and "owning actual Bitcoin," the fight I keep having with operators like Polto and Darko.

I had just interviewed David Bailey about feeding the incoming administration praise and money. It sounded to me like a protection racket. Lopp said the upside of Trump is that the man is incentives-driven, but "we've got Trump for four years and then that's it." He would rather not design Bitcoin around an election calendar. "I want to look at Bitcoin as a multi-generational project, not as a four-year election cycle project."

Stablecoins for Hotels, and a Short List of Non-Scams

Then I asked a sponsored question from SideShift: what does he do if someone offers to pay him in stablecoins?

"I actually get reimbursed in stablecoins for a lot of the conference travel that I do, and I highly prefer that to wire transfers. It's extremely convenient." He has nothing against them as a medium of exchange. He does not store his money in them. Would he swap them for bitcoin on SideShift? Usually not: he dumps the tokens for dollars to cover flights and hotels. His broader position is "to avoid having to sell Bitcoin at all cost." Most of his bills, and Casa's service providers, still only accept fiat. "We're still tethered to the fiat system."

I asked whether he had tried silent payments, which Cake Wallet supports (Cake was also sponsoring the episode). He had not. "I'm generally bearish on privacy on Bitcoin. I mean, I'm generally bearish on privacy everywhere, because so few people care about it." Privacy tools live on network effects, so low adoption almost defeats them. That is the same wall I keep running into in Bitcoin's privacy limitations. If the alternative is publishing your UTXOs to the world, he said, just accept Lightning.

BitVM, which Robin Linus kicked into public view the previous year, was another workaround that would work better if some opcodes were re-enabled. I asked whether rollups had dethroned sidechains, naming bridges such as BOB, Citrea, and Alpen Labs. He refused the verdict. Today's sidechains are generally federations, and a rollup is only as trustworthy as its sequencer. "Sidechains, yeah, we've never achieved I think what we really wanted to with them. Rollups aren't perfect either."

Then I asked him to name three coins he does not consider scams. "I've always been a fan of Monero. It does what it says it does." Grin would trigger people, "but Grin's basically dead at this point." Mimblewimble extension blocks on Litecoin work, even if they are not perfect. "Other than that, I mostly use stablecoins. Tether's the best shitcoin." After a pause he added Litecoin itself: not a scam, not pre-mined, and Charlie Lee dumped his stack. "Litecoin as a test network with actual economic value I think continues to be an interesting tool to just keep in our back pocket."

He did not name Zcash. I am mentioning the absence because privacy-coin conversations on this show often turn into a roll call, and this one did not. Andreas Antonopoulos had a presentation years ago about how every technology fails to scale by hitting a wall, then inventing a way past it. "The only problem is if we ossify and we literally cannot continue innovating. Then my fear is that the way that we scale is by everybody basically having IOUs at trusted third parties."

What Counts as Spam When Block Space Has a Price

I asked about Ordinals, since they had been the ones paying fees. He does not invest in them or trade them, and he is not against using Bitcoin block space for non-financial things, because he does not think you can feasibly stop it.

The spam argument, for him, splits in two. One definition is subjective. The other is economic. "Spam is just stuff that is not really economical for people to pay much for." People may dislike inscriptions, "but I think if inscriptions and the people behind them who are minting tokens and doing whatever they want on the network, if they're willing to pay more to use Bitcoin block space than people are to make an on-chain transaction, then they should get prioritized for that."

Peter Todd had gone to a Monero conference, come home, and written about tail emission: a small ongoing issuance that, if enough coins are lost, might not even feel like inflation for a century. Would that be the end of Bitcoin? Is Bitcoin just the 21 million coins?

"I think that's problematic, because 21 million is, I would say, one of the strongest and most fundamental memes and rules of the protocol." Tail emission is one option. He would rather explore many others, and he is biased toward the block space economics research he has been presenting.

I admitted I would get killed for saying this, then said it anyway. The Bitcoin Cash people were not idiots when they capped blocks at 32 megabytes. I had that conversation with Amaury Séchet, who treated the limit like a speed limit on a road. With Starlink and much better computers than seven years ago, is it a better time to raise the block size? Lopp called the hardware trend a truism. Moore's law, Nielsen's law, and Kryder's law all point the same way: computational resources get cheaper, so a node that costs about the same should be able to process more data. Bandwidth is the constraint he worries about most. If the block size never changes, "it's pretty much guaranteed that the cost of running a node will continue to go down."

He publishes annual Bitcoin Core sync benchmarks on a monster machine. I mentioned Eric Voskuil syncing libbitcoin on a cheap computer in a couple of hours. Lopp has not run that v4 architecture himself; as he recalls it, libbitcoin skips a separate UTXO structure and downloads more greedily from peers. Zero-knowledge proofs that let a mobile wallet sync in seconds are another option, "for some reason we're not there yet."

The Poster Boy Who Cooled Off After the War

There is a legendary photograph of Lopp in the Make Bitcoin Great Again hat, with the guns, the image people reach for when they want to explain what a bitcoiner is supposed to look like. He does not speak like the mascot, and he does not treat every proposal to change Bitcoin as a scam.

"I think a lot of that cultural shift happened during the scaling wars, and I was definitely a lot more toxic during the scaling wars, and I felt like it was appropriate at the time because we were basically in an existential crisis. And then I kind of cooled off, became a bit more moderate after that had all died down." What it comes down to, he said, is that he is a technologist who wants more experiments. "I think it's fundamentally kind of against the nature of how Bitcoin itself came into existence to be anti trying new things. Bitcoin didn't just appear out of thin air."

The experiment he had run most recently was the mempool research he presented at the conference that same day. On his talk list the NiceHash X slot is "The Prevalence of Private Transaction Mining."

I asked if he had heard of Hodling.ch, a Swiss shop that is a bit like Casa except they do not hold your keys. He had not. He knows the broader market of people who help you set up Sparrow or Electrum. Casa does that consulting too, plus its own software. "I'm not against people helping other people get into self-custody."

He does not have one hardware wallet that wins every category. Favorite metal backups are simple single steel plates. He is a fan of the Trezor Safe 3 and Safe 5 because they signed large, complex transactions in about a minute, whereas he has seen Ledger devices take hours. "If you want the hardcore crypto anarchist wallet, then you're probably going to go a Coldcard." None of them is best at everything, which is why Casa recommends multi-vendor multisig. I unloaded my Coldcard complaint: they talk like an open-source project, then fence the code so you cannot sell a fork. He let the rant stand.

Shor's Algorithm Meets a Three-Year Upgrade Lag

What he wants people to watch is quantum computing. Institutions such as the National Institute of Standards and Technology say we should prepare for machines that can run Shor's algorithm, derive a private key from a public key, and break Bitcoin's elliptic-curve signatures. People talk about a ten-to-twenty-year timeline.

He started talking about it because Bitcoin is in an ossification mood, consensus is slow, and he ran the numbers on how much block space a post-quantum migration would take: "that would be years." Then comes the uglier argument, about coins already sitting in quantum-vulnerable UTXOs. He can only see two choices: do nothing and accept that attackers will eventually steal that money, or make those outputs unspendable after a deadline.

"Either one of those breaks some important properties that a lot of bitcoiners hold very dear." Stolen coins break "your keys, your coins." Frozen coins violate sovereignty. I said that would be our Ethereum DAO moment. "Could be," he said. I raised the timestamping process as a reason a hard fork is inevitable. He was not impressed by that horizon. "That's not for decades, and I think that we're going to have critical issues come up long before then."

Is a hard fork still taboo? That question loops back to ossification, because "there is no authority or controller or even way to reach out and communicate with everyone who is using Bitcoin." He has been looking at how slowly node operators upgrade. "Right now I think it takes about three years for 95% of node operators to upgrade off of any given version of Bitcoin Core." A non-controversial hard fork is still feasible if it is scheduled far enough out, five or six or ten years, because running a node more than about two years old is already a bad security idea. If someone watching all of this feels jaded, he would point them at unfinished machines rather than a single hobby: e-cash mints and Fedimint, stealth addresses, and PayJoin, "a coinjoin protocol that I wish would be adopted more widely." His larger invitation: "This is a permissionless environment, so if there is something that you think is missing, then you can certainly try to build it."

I thanked him and said I hoped we would do this again, and that some of his assumptions would turn out to have been right. "I hope so," he said.

The through-line from S14 E9 is still self-custody. What this tape adds is the cost of treating Bitcoin as a finished statue while the largest balances migrate into a few vaults. Lopp will wear the paranoid crypto anarchist label that Saylor tried to make sound like an insult. He will not wear the other costume, the one that says the protocol is already perfect and the only remaining work is to cheer for the price. Bitcoin came out of people who kept trying things, the same refusal to treat a protocol as finished that I later unpacked with Perry E. Metzger.