The Throughput Basin: Cross-Substrate Convergence and Decomposition of Serial Decoding Throughput

Key Points

• This research aims to explore the convergence of throughput across various domains and develop a decomposition model.
• Examined throughput across six domains: genetics, artificial intelligence, human cognition, music, language, and engineered encodings.
• Analyzed 31 systems with varied vocabulary sizes and established effective information metrics.
• Developed the Throughput Basin Decomposition model to quantify information per serial event.
• Found effective information per serial event ranges from 3 to 6 bits.
• Convergence points in evolutionary simulations were identified at K ≈ 19–30 and ~4.1–4.6 bits.
• Statistical significance of throughput convergence is influenced by the inclusion of AI data.

Abstract