Gutenberg Lab — Morning Paper

The Day AI and Quantum Computers Break Blockchain Cryptography—Can Crypto Survive?

Speakers

Laura Shin (Host) × Connor Deegan (Co-founder & CTO, Project Eleven) × Alex Pruden (Co-founder & CEO, Project Eleven)

Bottom line

The threat of cryptographic breaks is materializing through AI faster than quantum computing, and the blockchain industry must migrate to "crypto agility"—supporting multiple cryptographic schemes simultaneously—immediately, or face catastrophic asset loss within years.

3-Line Summary

• AI has already broken experimental cryptographic schemes, making "AI Day" potentially more immediate than the quantum threat (early 2030s)
• ECDSA (Elliptic Curve Digital Signature Algorithm), used by Bitcoin and Ethereum, will definitely break under quantum computing, but Ethereum researcher Justin Drake warns AI could break it within months to years
• Unlike banks, blockchains cannot recover through identity verification if cryptography fails—they would become "graffiti walls," making multi-scheme support the only survival strategy

Three Key Points

1. AI is already breaking cryptography today: Two schemes (Hawk and McEliece) fell to AI attacks in the last two months. Anthropic broke one for a $100,000 bounty. While these were experimental schemes not yet in production, they were "candidates for deployment within a couple of years." Against the quantum threat timeline of 2030-2035, Connor Deegan emphasizes "AI Day is here now."

2. Elliptic curve cryptography is validated by a "trillion-dollar honeypot": The combined market cap of Bitcoin and Ethereum economically validates elliptic curve security. "If you break this, you win the jackpot and can take the money," yet it has stood for 30-40 years. While multiple hash functions have been retired, elliptic curves never have—this empirical track record is why Coinbase's Yehuda Lindell defends them.

3. Hash function randomness is unprovable: When asked "how do you actually know what's actually random," Connor admitted: "That is an inherently impossible question to answer. Cryptographers eventually just assume it's random." This is "one of the deepest questions of cryptography" and where hash functions have historically broken. The security foundation is faith-based, not mathematically proven.

Editorial Perspective

This debate reveals the tension between theoretical vulnerability and economic validation. Elliptic curve cryptography, though theoretically fragile, has withstood a trillion-dollar attack incentive. Hash functions, supposedly AI-resistant due to "lack of structure," rest on an unprovable assumption of randomness. What blockchains face is not choosing which assumption to trust, but whether they can build redundancy to survive any assumption breaking—a fundamental design shift required now. Unlike banks that can recover through identity verification, blockchains have no last resort. The reason "doing nothing" is the worst choice is not about technical correctness, but this asymmetry.

Source: Unchained "How Crypto Can Survive If AI or Quantum Computers Break Its Cryptography" (October 21, 2026)

https://unchainedcrypto.com

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --

OpenAI solves major math problems overnight, leaving researchers saying "I can't sleep"

Featuring

David Pierce × Robert Hart (The Verge's AI and mathematics reporter)

Bottom line

The speed and scale at which AI labs are solving unsolved math problems are invalidating research programs built over decades overnight, forcing the field to redefine its value from "producing answers" to "deepening understanding."

Three-line summary

OpenAI solved a Millennium Prize problem in 88 hours and released nearly 400 solved problems across more than 700 preprints in one evening. Mathematicians fear "the eye of Sauron" turning to their subfields, with some questioning whether to continue their research. This clash reveals a fundamental conflict between Silicon Valley's "move fast and break things" culture and academia's understanding-focused tradition.

Three key points

1. 88 hours, tens of millions of dollars to scoop competitors: When OpenAI heard a rumor that researchers were close to solving a Millennium Prize problem, they spun up ~10,000 agents, spent tens of millions of dollars, and solved it in 88 hours. Hart notes "the only reason they did it was someone else was [close]."

2. 400 problems, 700+ preprints dumped overnight: OpenAI's initial announcement of "more than 100" solved problems turned out to be nearly 400 results across more than 700 preprints released in one evening. Mathematicians say "it'll take years at least" just to verify them on a basic level, and three papers have already been retracted for "a sign error."

3. Researchers' existential crisis: Hart received messages from mathematicians saying "I'm struggling to sleep" or questioning whether to continue their research. Some had their "research programs blown up overnight," and the question is whether the profession can adapt before an entire generation of postdocs and PhD students lose their problems.

Editorial perspective

What's happening in mathematics isn't a race to produce answers faster—it's a fundamental question about how we measure the value of knowledge. OpenAI demonstrates the fact of "having solved" problems, but what mathematicians seek is understanding *why* those methods work, which becomes the foundation for the next breakthrough. This conflict will repeat across every academic discipline. While Silicon Valley chases quantifiable achievements, academia tries to preserve qualitative understanding—how we resolve this tension will shape knowledge production in the AI era.

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

Source: The Vergecast "AI is dividing the math world in two" (October 20, 2026)

https://www.theverge.com/the-vergecast

Can distributed energy scale through US utilities, or does it require market structure reform?

Participants

David Roberts (Volts podcast host) × Nick Chasseut (CEO of Octopus Energy US and Uplight)

Bottom line

Scaling distributed energy through utility partnerships is the realistic path in the US, but achieving UK-level results requires fundamental market structure reforms.

Three-line summary

Octopus Energy, which enrolled 60% of UK households in flexibility programs, has been confined to Texas retail after six years in the US and acquired majority stake in Uplight (serving 85 utilities) in September 2026. US utility programs achieve only 5-20% participation, and Octopus is testing whether it can reach UK levels through utility partnerships. The keys are overcoming three structural barriers: meter data access, billing relationship ownership, and residential battery export value realization.

Three key points

1. Texas price volatility "essentially destroyed"

Over the past 12-18 months, mass deployment of utility-scale solar and storage eliminated intraday price volatility in the energy-only market. This destroyed the price arbitrage opportunities that flexibility businesses depend on, forcing Octopus to pivot to standard retail growth. This exemplifies the fundamental paradox of flexibility business: success eliminates its own revenue source.

2. Residential batteries face 4x value mismatch

A residential battery might be 3-7 kW, but the average home peaks at 2 kW, so under current rules you can only realize about 2.5 kW of coincident demand reduction value. The battery can deliver 10 kW of load flexibility but only capture 2.5 kW of value, making investment hard to justify. In the UK, suppliers can fully realize flexibility value in their cost stack, so this problem doesn't exist.

3. Legacy billing systems are "single greatest hindrance to energy transition"

Chasseut identifies utilities' legacy billing systems (SAP, Oracle) as the fundamental blocker to flexibility at scale. Modern billing (Kraken) enables spinning up new dynamic rates in days to respond to distribution constraints; legacy systems take 18 months or literally cannot add new rates without deprecating old ones (West Coast utility example). Only one North American utility has adopted Kraken as of October 2026.

Editorial perspective

Octopus's UK success rests on three institutional pillars: forced choice, robust flexibility markets, and real-time metering. In the US, utilities own the billing relationship, meter data access is restricted, and residential battery export value cannot be realized—so the same playbook doesn't work. The Uplight investment is an experiment in "can we achieve the same goals through utility partnerships," but without utilities ceding full ownership of customer engagement, significant participation gains will be difficult. The real question is whether the US can scale distributed energy without market structure reform.

Source: Volts "Can distributed energy work around American utilities?" (October 21, 2026)

https://www.volts.wtf/p/can-distributed-energy-work-around

AI Disclosure: Produced with AI assistance; facts and analysis reviewed by our editorial team.

• --