Self-taught AI engineer and amateur linguist Tom Di Mino of New York’s Hudson Valley has announced a potential breakthrough in historical linguistics, claiming to have deciphered Linear A, the undeciphered Bronze Age script of the Minoan civilization. Di Mino, who initiated his research in January this year, identified a critical linguistic pattern on May 22 that he believes unlocks the syllabic system, which remained a mystery for over a century. His methodology and findings are currently under peer review by experts at Rutgers University and the University of Cambridge. Di Mino’s hypothesis posits that Linear A does not represent a Minoan language, as traditionally assumed, but rather an extinct Semitic tongue that served as a precursor to biblical Hebrew. To validate this theory, he leveraged generative AI tools, specifically Claude Code, to architect a computational engine capable of systematically cataloging symbols, mapping syntactic patterns, and stress-testing transliteration hypotheses. This technological approach allowed him to bypass the manual bottlenecks that hindered previous cryptographic efforts. The breakthrough centers on the analysis of formulaic religious inscriptions. Di Mino isolated a recurring verb structure containing five established Linear B signs alongside a unique Linear A character, designated as symbol 301. By cross-referencing this character with Semitic linguistic roots, he isolated the verb root nawaya, meaning to dwell or build. This aligns directly with the triconsonantal N-W-Y morphology common across Akkadian, Phoenician, and early Hebrew. Further analysis revealed that the inscriptions function as ancient prayers addressed to a goddess, exhibiting structural and thematic parallels to later Hebrew liturgical texts. The claim carries significant historical weight. Linear A, used between 1800 and 1450 BCE, shares 60 core syllables and several logograms with Linear B, which Michael Ventris successfully deciphered as Mycenaean Greek in 1952. While the overlapping symbols provided phonetic anchors, the remaining 13 exclusive Linear A signs and the script’s original meaning resisted decryption. Previous scholars, including Cyrus Gordon in 1957, suggested a Semitic connection but failed to produce a comprehensive translational framework. Di Mino’s AI-augmented statistical and grammatical analysis reportedly achieves what earlier manual methods could not, while also offering corrections to certain established Linear B translations, providing an independent validation of his model. If peer review confirms Di Mino’s methodology, the implications extend beyond ancient Near Eastern studies. The successful application of AI-driven pattern recognition to a complex, pre-alphabetic script demonstrates a scalable framework for decoding other undeciphered historical writing systems. Researchers at leading academic institutions are currently evaluating the algorithmic architecture, the proposed grammatical rules, and the resulting corpus of translated tablets to determine whether this represents a genuine paradigm shift in cryptographic linguistics.