During the Bitcoin Rails podcast #38 with Isabel Foxen Duke, Martin Shkreli opened by separating quantum reality from hype: he said quantum won’t replace Nvidia-style classical compute, but “when it comes to just Shor’s algorithm and just Bitcoin, you…have something to worry about.” The caveat: practical quantum machines remain slow, noisy and numerically fragile, and any credible attack would require error rates orders of magnitude better than what’s been demonstrated.

The core issue is fidelity. In quantum circuits, each logic operation (a “gate”) succeeds with some probability. Shkreli noted best-in-class gate fidelities around “99.99,” which sounds impressive until you compound the error across the “millions of gates” a full Shor run would require—at which point accuracy collapses. The upshot: either build vastly cleaner physical qubits or layer powerful error-correction codes to create highly reliable “logical” qubits. Both are hard.

He flagged IBM’s publicly accessible ~150-qubit systems as a reality check for enthusiasts: they’re educational but nowhere near the logical-qubit counts needed for a Bitcoin-scale break. Using the 256-bit curve as a yardstick, he discussed ballpark resource needs that reach roughly a million logical qubits—implying hundreds of millions to a billion physical qubits, depending on error-correction overhead. Today’s machines are orders of magnitude smaller.

Fresh out of prison in 2022 after being convicted of securities fraud, Shkreli dove into crypto, zeroing in on decentralized finance (defi) and the nuts-and-bolts of blockchain. During a X Spaces chat, he said he’d been using Uniswap—Ethereum’s leading decentralized exchange (DEX)—from behind bars, captivated by how it sidesteps old-school financial gatekeepers. Shkreli talked up Ethereum, Solana and Algorand, predicting ether could eventually overtake bitcoin by market value—a scenario often dubbed “the Flippening.”

During the interview, Shkreli noted that even with better physical qubits (he mentioned fluxonium as one contender), quantum teams must add many more “nines” of reliability. Shkreli contrasted classical graphics processing units (GPUs), which operate at extremely low error rates, with quantum hardware that’s still wrestling with noise, decoherence, and even cosmic-ray-induced bit flips. Until error correction truly tames those issues, Shor’s remains a whiteboard triumph more than a production tool.

Shkreli also pressed a crucial nuance: quantum isn’t “fast” in clock speed—often measured in kilohertz or worse—but valuable because certain algorithms (like Shor’s for factoring and discrete logs) change the math from exponential time to polynomial time. That complexity drop is the point; the hardware just isn’t there yet.

On timelines, he sidestepped hard predictions, acknowledging that credible Shor-class attacks on Bitcoin’s curve are not a five-year story and could take decades, particularly given the gulf between today’s physical qubits and tomorrow’s error-corrected logical fleets.

He did concede that non-quantum routes—mathematical breakthroughs potentially aided by AI—can’t be ruled out, but he still ranks quantum as the likelier first mover against elliptic curve cryptography. Either way, he framed the target as the cryptography, not “putting Nvidia out of business.”

Ethics surfaced, too: asked about “hacking Satoshi’s coins,” Shkreli said the intellectual achievement is the draw; he “wouldn’t want to hold those tokens,” calling that theft, even if the math succeeded. The research, he suggested, could be published without looting anyone’s wallet.

• What exactly does Shor’s algorithm threaten? It targets the hard math (factoring/discrete logs) behind elliptic curve cryptography used by bitcoin wallets, if a large, error-corrected quantum computer exists.
• Why aren’t today’s quantum machines a risk? Error rates compound across “millions of gates,” and current fidelities fall short for a full Shor run.
• How many qubits would an attack need? Shkreli discussed roughly a million logical qubits—implying hundreds of millions to a billion physical qubits with error correction.
• Could AI break bitcoin’s crypto first? Shkreli says advances in mathematics aided by AI are possible, but he still views quantum as the likelier first mover against ECC.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。