Crypto's Quantum Defense Is Being Built in the Open

What it claims

Eigen Labs' Yukon (launched August) runs open challenges in which developers and AI agents iteratively improve a benchmarked technical problem. The page draws a line between production deployments and benchmark-only results, and the summary blurs that line:

• Deployed or merged:
• Lighter.fast optimizations are in production.
• MLX.fast speedups are in Eigen's private inference network, Darkbloom.
• SNARK.fast optimizations are in Succinct's codebase.
• Benchmark or estimate only:
• ECDSA.fail
• HashSmash
• sig.golf
• better.codes
• The $59 Quantum Safe Bitcoin figure

How it works

1. A problem is opened publicly with an automated verifier.
2. Participants, alone or with agents, submit solutions.
3. Each submission is tested automatically.
4. An accepted improvement becomes the new baseline.

flowchart LR
  A[Public baseline] --> B[Human/agent submission]
  B --> C{Automated benchmark}
  C -- improves --> D[New baseline]
  D --> B
  C -- no gain --> B

ECDSA.fail started from Google's published verifier. Google released its quantum resource estimates and a way to verify them, but withheld the optimized circuits. ECDSA.fail asks participants to design circuits for one crucial step of a quantum attack on Bitcoin and Ethereum ECDSA.

Evidence

Two more details:

• sig.golf targets hash-based signatures that must also carry a mathematical security proof. The goal is to replace quantum-vulnerable transaction signatures.
• Quantum Safe Bitcoin protects transfers without changing Bitcoin's consensus rules.

How to use it

The page gives no commands or APIs. It states that HashSmash results are publicly reviewable and that Yukon accepts new challenge proposals.

Limits

• ECDSA.fail is theoretical. No existing quantum computer has demonstrated an attack at this scale.
• HashSmash (launched October 5 with Zcash, Shielded Labs and Project Eleven) has so far accepted attacks only on simplified SHA-256 variants, not the full algorithm. More ambitious submissions are still under review.
• The 10x Lighter figure is a benchmark result. Production saw a 2x reduction in proving time.
• The $59 Quantum Safe Bitcoin figure has not been demonstrated in a live transaction.
• better.codes remains well short of its 128-bit target.
• Source framing: this is a promotional Bankless piece by an author who has covered Yukon before. Benchmark definitions and verifier details are not given, so check each challenge's own page before relying on a number.