Show HN: SigRank – Competitive Stat Screen and Operator Performance Evals O7 (github.com)

SigRank has launched a privacy-preserving leaderboard designed to measure and rank the efficiency of AI operators based on their token cascade performance. Unlike traditional platforms that typically incentivize high-volume submissions, SigRank focuses on the concept of "yield," evaluating how much reusable signal is generated per input token rather than sheer quantity. The scoring formula is expressed as Υ = (cache_read × output) / input², enabling operators to understand the effectiveness of their AI sessions, emphasizing quality over quantity. This initiative is significant for the AI/ML community as it encourages operators to optimize their systems for efficient signal processing, potentially leading to more refined and impactful AI models. By providing an anonymous leaderboard that ranks operators globally based on their cascade shapes—classifying them into distinct signal classes such as Transmitter or Architect—SigRank promotes a culture of accountability and performance improvement. The platform's architecture, built using Next.js and TypeScript, ensures a user-friendly experience while safeguarding sensitive session data, as it relies solely on token-based metrics without revealing any content from user interactions.