EigenLayer

Yukon, Eigen Labs' open research platform, launched a new frontier challenge on the Dihedral Coset Problem (DCP) at yukon.org/dcp. Solvers must find the hidden reflection of a regular polygon from a set of quantum states faster than anyone currently knows how — a problem whose efficient solution would have direct implications for post-quantum cryptography, where DCP hardness is a core assumption.

• The launch extends Yukon's series of open, agent-friendly research challenges alongside HashSmash, CurveSmash, Labinius, and Circuit Breaker.
• Eigen Labs' EigenCloud account amplified the announcement, with Yukon tying DCP to Shor's algorithm, which lets a sufficiently large quantum computer break the elliptic curve cryptography behind ECDSA and RSA.

The Ethereum Foundation has opened evm.fast, an open challenge asking contributors to prove — with a machine-checked proof in Lean 4 or Rocq — a finite worst-case limit on how much computation it takes to verify an Ethereum block on a 64-bit RISC-V model. Today no one can prove a uniform ceiling, so the baseline score is infinity; the first finite, formally verified bound beats the record, and the real contest is then tightening that bound (measured as K, a certified upper limit on execution steps) toward a cost provers can plan around.

The model is compiled via Flapjack, a Lean 4 adaptation of the Pancake compiler, and startup steps count toward the total. Eigen Labs' EigenCloud account amplified the launch, arguing verifiable compute only works when the cost of checking is countable, with the challenge's repos living under Layr-Labs on GitHub.

• A certified worst-case bound would give real-time zk provers a math-backed budget and harden block verification against denial-of-service.
• No monetary prizes, deadlines, or submission platforms were announced at launch.

Circuit Breaker is live on YukonResearch, Eigen Labs' platform for open frontier research. The challenge asks solvers to prove a stronger unconditional circuit lower bound for one fixed, uniformly defined total Boolean family in NP — a long-standing complexity theory problem, seeded from a published 31/10 literature baseline rather than guaranteeing acceptance.

• Scoring uses joint coefficients and exponents, both maximized, with scientific-eligibility rules, machine-checked submissions, and explicit endpoints.
• Eigen Labs' EigenCloud account announced the launch, framing it as the next step in letting agents work in the open: every gain stays visible and verifyable, and the next solver can build on it.
• It extends Yukon's series of open, agent-friendly research challenges, alongside HashSmash, Labinius, and CurveSmash.

Yukon's HashSmash open challenge is live, built in collaboration with Zcash, ShieldedLabs, projecteleven, and Zooko Wilcox. The task: point thousands of AI agents at the hash functions that secure the internet — SHA-256, SHA-3, BLAKE3, and Poseidon — and see if they can find real attacks.

Early results landed quickly: researcher Dariia Porechna (@FutureDies) posted a collision attack on reduced 39-round SHA-256 that improves prior art by over 30 bits, plus what she called the first differential attack on that round better than generic search. Within days of it being posted, subsequent submissions on the platform made the attack 90x cheaper to run, and every result stays open for the next agent to build on. Zcash co-founder Zooko Wilcox reported multiple attempts within the first 23 minutes of launch.

• All results are reviewed and left public, keeping the research agent-friendly and open.
• The claim, framed as SHA-256 "not broken" but its security margin "lost 1 more round this year," was amplified via Eigen's cloud account.

CurveSmash is live on YukonResearch, an open cryptanalysis challenge run in partnership between Aztec and Eigen Labs. Aztec co-founder Zac Williams is leading the new open frontier challenge, which Eigen Labs founder Sreeram Kannan framed as putting frontier attacks on elliptic curve cryptography in public — the launch was amplified by Eigen's cloud account. It tests the hardness assumptions of elliptic curves, the math a large part of the internet runs on.

The challenge puts AI agents to work attacking secp256k1 discrete logarithms (the curve behind Bitcoin and Ethereum) and BLS12-381 fixed-generator pairing inversion using full production parameters, on the premise that quantum computers will eventually break them, so the public should learn how much time is left before anyone else does.

• Participants compete on a single metric — log2 of total computation in target-group-operation equivalents, lower is better — with preprocessing and failed attempts counted, and cryptographers reviewing every submitted result.
• CurveSmash follows HashSmash's exploratory scoring standard, where a concrete supported claim can score when no fatal flaw survives defense and adjudication, extending Yukon's series of agent-friendly cryptography challenges.
• Aztec's David Steinrück celebrated the launch with a pun on the target curve, writing "everything is breaking, but we're ahead of the curve (secp256k1 that is)."

Eigen Labs and its Yukon platform launched a new open challenge called Labinius, focused on lattice-based proof optimization for post-quantum cryptography. The challenge aims to move past security alone and drive the performance that makes proof systems practical to deploy, with both researchers and AI agents able to contribute measurable improvements.

• The challenge was announced live via the EigenCloud account on YukonResearch.
• It extends Yukon's series of open, agent-friendly cryptography challenges, alongside the earlier hash-function and post-quantum signature rounds.

PrismML released Bonsai 2 in three new sizes — 2B, 4B, and 9B — open-weight ternary models built on Qwen3.5 and sized for everyday devices (2B at 0.43 GB, 4B at 0.95 GB, 9B at 2.0 GB). The 2B reaches 90 tokens/sec on an iPhone 17 Pro Max with reasoning, vision, and tool calling running entirely on-device.

Eigen Labs' EigenCloud account amplified the launch by retweeting the release announcement, pointing to Bonsai 2 9B available to try on Darkbloom and tying the release into Eigen's push for open, compact AI models.

Darkbloom's Bonsai 2 9B is live on the PrismML service via Yukon, compressing Qwen 27B to 8.5 GB while retaining 98.2% of its benchmark scores — small enough to run on a 16 GB Mac. Yukon, in partnership with Darkbloom, then opened an optimization competition asking the community to make the model faster, with Bonsai 2 9B packing reasoning, vision, and tool calling into a single compact model.

• The community has already cut inference time to 2.26x faster than the starting point.
• The optimization window stayed open until October 1.
• The competition was announced by PrismML and amplified via Eigen's cloud account.

Eigen Labs' EigenCloud account laid out the company's next chapter: an open AI economy that people can build on and earn from. Per the post, DarkbloomAI puts idle Macs to work serving paid inference, while YukonResearch brings humans and agents together to advance frontier research.

The message frames research contributions and paid inference as open, earnable infrastructure rather than closed products, extending Eigen's pitch of open, agent-friendly systems beyond Yukon's research competitions.

Eigen Labs restated the same pitch in a post pointing to Blockworks' reporting on crypto inference providers, noting that demand is outrunning supply despite trillions invested in compute globally. EigenCloud amplified that post too, tying Darkbloom and Yukon together as the infrastructure and research halves of the open AI economy.

Eigen Labs founder Sreeram Kannan announced a new open study on the Yukon platform, arguing that AI methods are now breaking long-standing cybersecurity assumptions and that it is time to re-evaluate the field's foundations. The announcement sets up a public, agent-friendly research effort on Yukon rather than a closed lab.

• The launch notes that turning a paper attack on SHA-256 into a real one took cryptographers 11 years — a problem being handed directly to AI agents.
• Comes as Yukon runs a series of cryptography-themed open challenges alongside StarkWare, the Ethereum Foundation, and AI-agent attack rounds on SHA-256, SHA-3, BLAKE3, and Poseidon.
• Amplified via Eigen's cloud account.