The study, titled 90 Years Engraved by Entropy, utilized the newly launched LingEQ Engine to evaluate texts ranging from 1936 to 2025. Shannon’s A Mathematical Theory of Communication achieved a Linguistic Entropy Quotient of 194, narrowly edging out Turing’s 1936 paper, On Computable Numbers, which scored 193. By contrast, modern breakthroughs like GPT-3, AlphaFold, and DeepSeek-R1 clustered significantly lower, between 168 and 170. Even the 1955 Dartmouth Proposal, which formally christened the field of artificial intelligence, registered a score of 150.

Researchers suggest these figures reflect a maturation of the field. While early AI papers forged entirely new theoretical boundaries, contemporary research largely operates within pre-existing paradigms. This analysis arrives as the scientific community struggles to filter the 20,000 preprints uploaded to arXiv monthly. Standard metrics like citation counts and peer consensus often fail to quantify the actual intellectual substance of a paper, favoring popularity over foundational depth. The LingEQ Engine, which began its global beta on August 1, 2026, aims to provide an objective, text-based alternative to traditional evaluation methods, with its full theoretical framework scheduled for publication by Springer Nature.