The Café Bitcoin Podcast

The last day. Brady opened with the count, 37 shows in 50 days, and the campaign's frame: political independence 250 years ago, monetary independence now. Yan Pritzker, Frank Corva, Suze and Alec on stage; Unseen Finance returned at the end.
The Liquid hack, plainly. On Sunday somebody sent 4,000 L-BTC to SideSwap's peg-out and the Liquid Federation paid out real Bitcoin. The L-BTC had been created from nothing. Blockstream paused the sidechain and 3,400 Bitcoin came back Monday.
⭐ Yan on what broke: a caching bug, not anyone's keys. "The only two hard problems in computer science are naming and cache invalidation." Two transactions shared one cache key, so validators mistook a fake for one already checked.
Why it is not Bitcoin. Yan: Liquid is an 11-of-15 federation of companies with its own database and rules. "This is a separate thing, this is not Bitcoin itself." Lightning is built on Bitcoin script; Liquid is not.
⭐ Frank on the 15%. The hackers returned 85% and kept about 600 Bitcoin. His question: will anyone be made whole, and if they keep it, "they're surely not white hats." A Bitcoin Policy Institute lawyer offered to talk.
Yan's verdict on the hat color: gray. Taking everything to beat a black hat is arguable. Keeping a cut is not. "You don't just return a wallet and take a bunch of stuff out of it."
Suze's question: Bitcoiners are ruthless about added trust assumptions, so why did Liquid get a pass? "Were we discerning enough?" Leave the base layer and you take on another code base and another set of failure points.
⭐ Yan on open source in the AI era: it stays because every open project gets red-teamed by the whole internet. Closed systems get hacked too, "we just don't hear about them." The base chain is "the ultimate boss" nobody has beaten.
Yan's bank numbers, read live from the Identity Theft Resource Center: 3,332 reported data compromises in 2025 affecting 280 million people, 739 of them in financial services. "Bank-grade security" is not a standard anyone actually meets.
⭐ Frank's hard question: must the layers be as bulletproof as Bitcoin? Yan: nothing is, and the only path is slow growth. Bitcoin was attacked at $1, then $10,000, then a billion. Early adopters carry the risk.
Yan's rocket-ship rule. Bitcoin Core engineering is closer to launching a spacecraft than shipping web software. "Years of careful review, study, simulations on other chains." Test new tech on other coins for two years first.
Suze asked whether ossification risks losing the skills to change Bitcoin in 50 years. Frank and Yan: no. Ossification is no new features, not no developers, and AI has widened who can work on cryptography and game theory.
Yan on open-weight versus hosted AI: the same debate as open source versus closed, and open source won the internet. His advice to companies: externalize memory, skills and harness so you can swap models. Swan built its own.
Unseen Finance returned with a proposal: these incidents are a credibility problem for the services on top of Bitcoin, not for Bitcoin. His fix, borrowed from banking: a self-regulatory body plus an insurance pool that service providers pay into.
Brady's tally at the close: Cory, Yan, Steve, Alec, Isaiah, Matt and Brandon from Swan; Larry Lepard three times; Parker Lewis, Daniel Batten, Tom and Callum, Guy Swann, Ryan MacLeod, Alana, Bob Burnett, Roxana. Suze made 35 of 37.
The close. "We have the tools, we have the knowledge, we have the culture, we have the determination." The half-price buy fee ends today; limit orders placed today keep it. Yan's last word: take security seriously, "go play with AI."

A Friday hang that took a turn. Tomer Strolight, Suze, Swan Private's Alec and Matt, then Brandon Quittem and Dirty Coin director Alana. The surprise came from the audience: an anonymous 40-year IMF and central-bank veteran, @UnseenFinance.
Brady opened on a viral Instagram video, described on air and unverified at air time: Meta AI's suggested questions on a mother's post surfaced her child's name, her location and a deleted photo. "Breadcrumbs for stalkers laid out nice and neat."
Tomer called it the new reality and a double-edged sword: the panopticon that erodes privacy also finds a missing child after an Amber Alert. Every recorded word now has "perfect recall," which he called "a metaphysical change in how reality works."
Brady's frame: decades of data on a public network, then phones with microphones and cameras, then AI to stitch it together. "Orwell was right. Maybe it didn't play out exactly the same way, but we all opted into it."
Suze went further: worse than Orwell, who never predicted programmable money. Her point was aggregation, a phone number, an employer, a public record: "once data gets copied, leaked, sold and ends up on the dark web, it is impossible to claw back."
⭐ The surprise guest: 40 years across the IMF, development banks and central banks, "what often Bitcoiners call the enemy." KYC scope creep brought him to Bitcoin: rules for big criminals now "attacking single mothers. How dare you brought in $300?"
Alana argued most people don't value privacy. Her rule, from a Spanish saying: closed mouths, flies don't fly in. Her forecast: "it's going to get worse before it gets worse," then regulation arrives "for the safety of the children."
⭐ Tomer pushed back on the doom: technology is unstoppable, so use it morally and defend yourself with it. Embracing coal led to gas and electricity; fearing nuclear for 70 to 80 years left the world with worse energy and worse governments.
⭐⭐ Suze asked the ex-IMF guest whether institutions understand the surveillance they build. His answer: the people at the top are excellent and well-intentioned; the damage happens in the "telephone game" downward, ending with a bank clerk blocking a payment from fear.
Suze quoted Carstens on CBDC control; the guest said central banks "100%" believe CBDCs expand access and reduce poverty. Brady disagreed: security, poverty and children are window dressing; the Patriot Act and KYC/AML show how narrow mandates become mass surveillance.
Brady's hope story: Snowden broke his faith in government; Bitcoin and the cypherpunks restored it. His ask now: Bitcoiners should learn open-weight AI models the way they learned nodes and wallets, as personal defence against deep fakes and scrapers.
⭐ Brandon brought Kevin Kelly's What Technology Wants: technology is an extension of evolution and every technology carries a bias. A CBDC wants control and reduced privacy; Bitcoin wants individual sovereignty. One panelist's summary: Promethean versus Faustian technology.
Brandon raised Gladstein's structural-adjustment critique. From the inside, the guest said, "that's not true": governments come to the IMF, then blame it for the conditions while "the IMF sits back and takes that black eye." Brandon: bad outcomes, bad program.
Tomer's reconciliation: there is no parallel universe to test the counterfactual, so fallible people with incentives argue over outcomes. Technology's biases compound objectively; human biases compound with incentives. He credited Erik Cason's "oath of the machines" from Cryptosovereignty.
The guest closed on a "eulogy for Bitcoin" he published in January: too much price pumping and Wall Street, contrary to self-sovereignty. Tomer's counter: the players who exploited fiat are arriving, but "bailouts are still impossible in Bitcoin."
Close: no show Monday, the finale is Tuesday, September 8, the last day of the half-price buy fee. Brady thanked the surprise guest: "it's fun when we do these spaces to get surprises like that from the audience."

A mining show, and the Space crashed midway. Bob Burnett, elected chairman of OCEAN two days earlier, joined Tom French and Callum Wheeler of RenewaBlox, with Suze and Swan's Alec on stage.
⭐ Burnett's hashrate taxonomy now has three animals: rabbits (home miners, maybe 1 to 2% of hash), wild horses (miners who produce their own power), and captive elephants (industrial miners on someone else's grid).
His definition of the risk: "captive means you are dependent on somebody else to provide you with the power. Wild means you produce your own energy, and so therefore you have freedom."
⭐⭐ His target, said plainly: if wild horses reach roughly a third of the network, "we have a great deal of resistance to hash centralization." His own read is that the elephants have shrunk as public miners pivoted toward AI.
⭐⭐ Asked whether mining stabilizes the grid, he conceded what the industry rarely does: "we overstate our ability to have impact." Bitcoin uses half a percent of world electricity, and "we cannot stabilize the world's grids. We're not big enough."
The scale he put on it: roughly 160 terawatt-hours a year against about 30,000 globally, and a block-reward industry of "roughly $10 billion" in 2026. Local and regional cooperation, yes. Macro stabilization, no.
Barefoot Mining's own model is the wild horse in practice. The company owns gas wells in Pennsylvania, runs engines Burnett says burn nearly emission-free, and is investing heavily in stranded natural gas over the next six to twelve months.
⭐ RenewaBlox brought the UK version. Callum described two run-of-river hydro projects in Scotland, one roughly a megawatt and fully off-grid after the network operator made a connection uneconomic.
⭐ Tom's small-site example was the one that made the economics concrete: a 200-kilowatt hydro site with export capped at 150, leaving 50 kilowatts of headroom and, by his figures, around 18 to 19 pence per kilowatt-hour on the table.
The Space dropped right there, mid-sentence. The recording picks up with "Are we back?" and the room reassembled within a minute.
The heat lane: a distillery in Annandale is piloting a sand battery, turning excess electricity into stored heat through a resistor. RenewaBlox's first thought was that the resistor could be an ASIC.
Suze relayed a listener's challenge that Burnett had said mining doesn't stabilize the grid. He confirmed it, which is how the concession above got made.
Tom on where efficiency gains come from: today's rooftop solar is bounded by the Shockley-Queisser limit, and stacking perovskite tandem layers could lift panel output roughly 10% per layer, with five or six layers as a plausible ceiling.
Burnett on AI: he uses it daily to run operations and won't put constraints on what it might unlock, but for his lifetime the energy story stays gas and renewables.
The campaign is in its last days. No show Monday. Tomorrow is the last episode of the week, and the final episode is Tuesday, September 8, also the last day for the half-price buy fee.

Brandon Quittem walked through an essay he has never published, and Brady said so up front. The idea: Bitcoin resembles the Apache, who held off three empires for centuries because there was no leader to remove.

He opened with the disclaimer, and it matters: "analogy, analogy, analogy. It's not one to one. These are just frameworks."

⭐ The mechanism he laid out: a centralized organization has a choke point, so Cortés could take an empire by capturing one man. The Apache had no such point, so killing a leader just produced two bands instead of one.

⭐⭐ What finally worked was a gift, not a fight. The US gave Apache leaders 100 head of cattle. A people who had never stored wealth formed a council to divide it, and the wealth eroded them from inside.

His one-line version of the lesson: "you cannot fight decentralized organizations face-on. You have to subvert them from the inside, through values, through corruption, or by applying leverage on leaders."

⭐ He called it the golden cow, and named the Bitcoin version carefully: early holders get wealthy, stop fighting, and the supply drifts into a few custodial products. Then he hedged: he is not claiming we are there.

Brady asked whether the cattle were a knowing move by the US government or an accident that happened to work. Brandon treated it as an open historical question rather than asserting intent.

Brady's larger question ran under the whole hour: does culture even matter to Bitcoin? Brandon traced the waves, cypherpunks first, then libertarians and sound-money people, then the writers, and argued each cohort changed what the network was defended for.

A listener's take on custody was the practical takeaway: custodians have gotten much better, and the thing to protect is not a purity test but the standing ability to exit to self-custody quickly if you ever need to.

⭐⭐ Then the fungi, and the origin story, told here for the first time. He wrote part one of the mycelium essay in one night after a meetup. Dan Held edited it and gave him confidence to publish in 2018.

He was gun-shy because the genre didn't exist yet: "in 2018 there wasn't a lot of esoteric Bitcoin-is-a-living-organism talk." Held's own four-part series had compared Bitcoin to a tree and Satoshi to a gardener.

⭐ The biology, in his words: mycelium is the organism and the mushroom is just the reproductive organ. The network has no brain, runs underground or inside trees, and moves information in both directions across long distances and across species.

Brady brought the pop-culture version: Star Trek Discovery built a spore drive on an intergalactic mycelial network, with an astromycologist character named Lieutenant Paul Stamets after the real mycologist Brandon cites.

Brandon on why fungal medicine works on humans: the biology is close enough that a molecule a fungus builds to fight bacteria often works in us. His estimate is that over half of modern pharmacology is fungal or fungi-derived.

The detail that got the room: Ötzi the Iceman, the frozen body found in the Italian Alps with an arrow in his back, was carrying several kinds of mushrooms when he died.

His stated principle for the whole project: a "polymathic responsibility" to combine what you know from elsewhere to explain a genuinely new thing.

Tomorrow: Bob Burnett, newly promoted to chairman of Ocean Mining, on what comes next for the pool.

No booked guest today, and the show was better for it. Brady opened the room to listener Bitcoin questions, and the format ran most of the hour.

⭐ Suze gave the single sharpest illustration of UK debanking anyone has offered on this show: Bitcoin Policy UK calls itself BP UK because naming Bitcoin or crypto makes it materially harder to obtain banking services.

The UK debanking inquiry's evidence window closed the day before. Suze confirmed no findings yet, the group collects every submission, then publishes a report.

She was precise about what the body can and can't do: an All-Party Parliamentary Group isn't made of ministers and holds no legislative power. It advises MPs and puts an issue on the table.

Her frame on the underlying problem: roughly 40% of these businesses face blocked or delayed lawful transactions, and firms are leaving the country. Her line was that it shouldn't take an inquiry to point out the obvious.

⭐ The longest conversation of the day started with a listener question about whether anything replaces the liquidity role stablecoins currently play, and ran deep into how Lightning channels, sidechains and settlement layers actually differ.

The practical answer the room landed on: liquidity isn't a technological hurdle, it follows demand. The reasons things settle where they settle are mostly regulatory and operational rather than technical.

⭐ Cory's case against Bitcoin price models: a best-fit line through past data has no predictive power, and stopping your sample earlier doesn't fix it, because dozens of lines fit any window.

He'd tested it before: using only 2010 to 2012 data he found lines that predicted Bitcoin's price through 2022 better than stock-to-flow did.

His verdict on the genre, and it's the keeper: the models smuggle in an assumption about future adoption and then present it as knowledge. "This is just astrology for dudes."

⭐⭐ The GBTC segment came with real numbers. Roughly $8.5 billion still sits in the trust at about six times the cost of competing products, much of it in brokerage accounts people haven't looked at in years.

⭐ And Swan is actively doing something about it. Alec has been running conversions out of GBTC into directly held Bitcoin through Real Bitcoin Exchange, without selling on the open market, and the original cost basis carries over.

⭐⭐ Brady closed on the tagline from his own podcast, and made the historical case for it. Money is the foundation of civilization, because everything a society builds that outlasts one lifetime requires somebody able to save.

Rome minted the denarius from 211 BC and held it for 275 years, roughly ten generations. The roads, the aqueducts, the law and a trade network from Britain to Syria were built on a coin that didn't move.

Then Nero clipped it in 64 AD, the year Rome burned, taking the silver to about 94%. Nobody rioted. Marcus Aurelius took it to 75%, Severus under 60%, and it reached roughly 2% within a century.

⭐ Four coins, three civilizations: denarius 275 years, solidus about 700, ducat over 500, florin 281. The flowering sits on top of the stable money; the decline begins after somebody starts clipping it.

Suze's lightest moment: she watched Hamilton with her daughter and realized partway through that it was the story Cory had been writing about all along.