S16 E44: Abdel on ZK STARKs & Bitcoin Maximalism

"A weird animal in this space." That is how Abdel describes himself, and after three hours and eighteen minutes live I have no better label for a StarkWare employee who spent years as an Ethereum core developer at ConsenSys, worked on Hyperledger Besu, and now calls himself a Bitcoin and Ethereum maximalist in the same breath.

I had him on because his résumé reads like a map of the arguments I keep having on this show. He championed the fee burn in Ethereum, left over a trusted setup, and now works on STARK proofs of the Bitcoin header chain, a Cashu fork, Nostr experiments, and research on quantum signature aggregation. He also had a number for me that I did not believe until he explained it: a proof covering sixteen years of Bitcoin headers, produced in 6.5 hours, for the price of a Big Mac combo.

The conversation wandered a lot. We covered rollups, OP_CAT, a viewer who rage quit, my bank's paperwork demands, and French suburbs. I will keep to the order that makes the best story rather than the order we recorded it.

Before Fighting Banks, I Was Working for Them

I started with a question I ask anyone who comes from fintech: does that background push you toward DeFi? Abdel said the opposite happened. "Before fighting banks, I was working for them," he told me, and that experience taught him how fiat rails behave when you are on the wrong side of them. So he cares more about systems nobody can stop than about yield farming.

He also dated his own awakening. Early on, he said, "I was not mature enough to understand the potential impact of this, especially on the political level and on the economic aspect." And he insisted the Ethereum label never covered all of him: "I never was only Ethereum," he said, adding that "I was describing myself as a Bitcoin and Ethereum maximalist."

That led me to my own bank story. After I bought a car, my bank demanded receipts for transactions from a decade earlier to prove where my Bitcoin came from. Abdel did not need convincing about where this goes. His view is that engineers get lost in the fun parts: "I think there is this technical distraction for engineers that like to work on cool stuff. But sometimes it's a distraction that can make you forget about why we are doing all of that."

He also thinks the whole system is too fragile if it leans on a few people. "We cannot rely on just a small group of technical people because they can easily be captured or put in jail and everything like that," he said. On the Ethereum side, he pointed at how ordinary the exposure has become: "Everyone literally is fine using MetaMask connected to a public RPC and no one gives a shit about that, which is very bad."

The Fee Burn and the Dangerous Precedent

We spent a good stretch on EIP-1559, because Abdel was in the middle of it. "It was EIP-1559, which was the introduction of the burn, the fee burn in Ethereum," he said, and then, without being asked, "I was the champion of it."

I framed it the way a Bitcoiner would, as a council of people deciding to change monetary policy. Abdel corrected the history. Vitalik proposed it, it was dismissed, and ConsenSys picked it up afterwards. It went through public polls and calls with miners, so in his telling it was far from a closed-room decision.

What he took from it is a warning, though. "The fact that we changed it on the fly is an interesting precedent," he said, and then: "if you can do it for a good thing, maybe you can do it also for something bad." A community that gets comfortable altering the rules when the change looks good has built the habit it will need when the change looks bad.

Jumping Out of the Pot

The reason he left Ethereum is a metaphor about a frog, and he told it well. The slow heating, the small compromises that each seem fine, and then the realization at the end: "At some point you wake up and you realize that you are too weak to get out of the water and you will just die."

The specific trigger was the trusted setup needed for danksharding. "I'm very anti-trusted-setup in general, that's why also I prefer STARKs over SNARKs," he said. And about the moment he left: "That's when I felt that the temperature was too high and I had to jump."

His ranking of risks for the two networks is lopsided. For Bitcoin: "The risk on Bitcoin is mostly on the tech front, like becoming obsolete and not fixing the stuff that we know have to be fixed, which are security budget, privacy, quantum resistance." For Ethereum, the problem sits in the people. "You can always fix a broken tech, but the broken social layer is forever."

A Rage Quit and the Weekend Pay-Per-View of Ideology

About ten minutes into the stream, a viewer rage quit and called the conversation "intellectual retardation." Abdel's response was one line: "He doesn't get the point."

The point, as I understood it from the rest of the night, is that Abdel believes you cannot be anti-Bitcoin if you understand what it does. "There is no way you can be anti-Bitcoin," he said. He also mentioned that he never bought into the narratives of the block size era: "I never bought into this bullshit," and of that period he said "I think it was the real revolution, to be honest."

I brought up a story from that same war. A Bitcoin Cash developer once discovered Matt Corallo's accidental inflation bug in BTC and disclosed it to Core developers, even though the two camps were at each other's throats. Abdel took that as part of the same picture: technical people sometimes do the right thing across enemy lines, and tribal noise hides it.

The tribalism bothers him on the other side too. "This is something I really don't like in this new wave of Bitcoiners that are just simping MicroStrategy, Saylor and ETFs. This is insane to me," he said. I added my own observation that the conversation has split into two camps: "It's sort of degenerated to the point where in one side you talk about gains and you talk like Wall Street people. And on the other side, there is actual cryptography talk."

STARKs, Verified in Milliseconds, Proven for a Big Mac

This was the technical center of the night. I asked him to explain the acronym, and he started with the letter I always forget. "The T is for transparent, which means that there is no need for a trusted setup." STARKs use hash functions instead of elliptic curve pairings, which is also why they are quantum-resistant.

Then the economics. "The magical property of STARKs is that it is exponentially faster to verify the proof as opposed to re-execute the original computation," he said, and "the more computation you aggregate into one proof, the more interesting it becomes."

Here is where the Big Mac comes in. StarkWare used its Stwo prover to generate a STARK proof for the entire 16-year Bitcoin header chain in 6.5 hours, and the cost was roughly a Big Mac combo. Abdel made sure I understood which side of the process that number belongs to: "the Big Mac is for the proving." Verification is the cheap part: "Anyone can verify it in a few milliseconds."

He thinks this distinction gets lost in Bitcoin circles. "There is a lot of misconceptions and confusion in the Bitcoin community about the distinction between proving and verification," he said. And his plain-language version of the whole idea: "STARK proof is just a new way to do that without even executing it. I want people to understand that it's a new way to verify the integrity of a computation without executing it."

The trust model is the part that excites him most. Compared with BitVM-style bridges, "you don't even have the one-of-n honest assumption. You need to trust no one."

On the quantum angle, he has been doing research on signature aggregation, and I pointed him toward people I have had on who work on that problem. Hunter Beast went through BIP 360 with me in S16 E35.

Cashu, Cairo and a Decentralized Amazon

Abdel's Cashu fork is where the STARK work turns into something a user could touch. Cashu is the ecash system for Bitcoin, and his version adds programmable privacy through proofs written in Cairo, the language behind Starknet. The spending conditions stay hidden from the mint. "The mint will just verify the proof, but he will not learn anything about the spending conditions," he said.

His second experiment is on Nostr. The idea is to use STARK proofs so that jobs done by anonymous workers can be checked without trusting them. He called the end state a "decentralized permissionless Amazon." I asked how that differs from the usual marketplace pitch, and his answer came down to verification: the buyer does not need to trust the seller's machine if the proof comes with the result.

He gave credit where he thought it was due, too: "He's literally dedicating his life to building freedom tech and privacy." That was about a builder in the Cashu and Nostr world who has stuck with it for years.

The Nostr experiment also intersects with a question I keep returning to on the show, which is how much computation Bitcoin should be expected to do natively. My conversation with Eli Ben-Sasson on STARKs, rollups and privacy is S15 E36.

Rollups, Citrea and the Duke Nukem Problem

I gave Abdel my usual provocation about ZK rollups on Bitcoin: the technology reminds me of Duke Nukem Forever, always promised and always late. He answered with market history. Optimistic rollups launched first, without working fraud proofs, and they still dominate. Shipping early with a weaker design beat waiting for the clean one.

He also supports the Bitcoin projects working in this space. "I'm a big fan of Citrea, by the way," he said, and he thinks the criticism of Citrea's use of OP_RETURN is unfair. The same goes for Alpen Labs, whose approach I went through with David Seroy in S16 E27. On where bridges are going, he pointed to "the future of BitVM bridges," with BitVM as the base layer for them.

On the technical merits he is not subtle: "Everyone knows that ZK rollups are way better on every level." Doing the same work with only today's Bitcoin script is possible but painful: "You can do it, but not in the most efficient manner or in the most elegant way."

We also compared notes on decentralization claims. Abdel argued that Starknet has more diversity among its sequencers than people assume. "In fact, we are literally the only layer 2 that has this diversity. In fact, we are even more decentralized than L1s, which is surprising, like more decentralized than Solana or any L1s," he said. I did not argue the point on air, so take it as his claim about his own project.

The $1 Million Fund and the Swiss Army Knife

The OP_CAT discussion is where I pushed hardest. StarkWare put $1 million into a research fund for Bitcoin covenant work, and I wanted to know how that squares with neutrality. Abdel described it as a neutral grant, meant to let critics demonstrate the harms they keep warning about, with money behind the demonstration. If the fear is that covenants will wreck Bitcoin, he reasoned, someone should be paid to show it.

On how to propose changes, he rejects the cautious route most people favor. The "smallest change" strategy for OP_CAT, in his view, is a trap. He prefers a bundle, a "Swiss Army Knife" package of OP_CAT, OP_CTV, OP_CHECKCONTRACTVERIFY and OP_AMOUNT.

He also told me about a moment that surprised him. Luke Dashjr tweeted that Bitcoin needs a ZK verifier. "I was extremely surprised by that," Abdel said, and he followed up with a prediction: "people will be forced in a way to have a position about that." That tweet made him think about an OP_STARKVERIFY opcode, an experiment that could settle the generic opcode debate by giving people one concrete thing to accept or refuse.

His argument against waiting is the part I keep chewing on. "Actually doing nothing is the biggest risk of capture that you can have," he said, meaning that institutions may eventually own enough supply to block any change at all. The window for fixing security budget, privacy and quantum resistance, in his reading, closes if we spend it arguing about process. "We need to tend towards ossification at some point, but not now."

I raised my own history with this. I once ran a Romanian gaming magazine experiment where readers could pay in BTC, and the friction of local Bitcoin ATMs forced customers to centralize orders through friends. It showed me how quickly a decentralized network pushes people toward trusted intermediaries when the tooling is thin.

Europe, Immigration and the Only Network With a Chance

Late in the night we drifted into politics, and Abdel was direct about it. "For me, like Europe is cooked right now because the problem is that it's a very bad cocktail because it's bad both economically speaking, innovation, tech and also kind of identity crisis," he said. We talked about French immigration and suburban integration. I told him about my semester in Villiers-le-Bel on a 500-euro scholarship, and he had his own perspective as someone who grew up in that country.

He ties it to money. "Bitcoin is also maybe a peaceful exit to the fiat system and how not having a hard money has a lot of consequences in the world even for wars and conflicts," he said. That fits with his earlier point about banks, and with why he thinks decentralization is easy to undervalue: "Decentralization and censorship resistance don't matter until they do."

I showed him the relic I had brought, a destroyed laptop that once ran a node, along with signed DVDs. "I bought this for like a thousand bucks from him in Bitcoin," I told the stream, about Peter Todd.

His case for Bitcoin over the alternatives is not sentimental. "I don't think it's possible ever to reproduce the conditions that were met for Bitcoin to be where it is today," he said. "I became a maximalist because I think it has the most chances to succeed and the chances are not that high." His timeline is long: "It is the early days. We are building stuff for decades to come."

The concrete plan he left me with is to keep pushing on three things before ossification: the security budget, privacy, and quantum resistance. His own work feeds directly into those, with the Stwo-proven header chain, the Cairo-based Cashu fork, the Nostr verification experiments, and the signature aggregation research. He plans to keep building all of it from inside StarkWare, and to keep arguing for the Swiss Army Knife package every time the opcode debate comes around.