Self‑taught AI engineer Tom Di Mino says he has cracked Linear A, the Bronze‑Age Minoan script that has defied scholars for 120 years. Linguists at Rutgers and Cambridge are reviewing his submission. If his solution holds, it would echo Michael Ventris’s 1952 deciphering of Linear B and could rewrite early Mediterranean linguistics in the field of archaeology.

Di Mino argues the symbols encode an extinct Semitic tongue, a precursor to biblical Hebrew. He fed a corpus of Linear A tablets into Claude, an AI code engine, to map recurring patterns. The breakthrough emerged on May 22 when he identified a five‑sign verb root “nawaya” meaning “to dwell,” linking the prayer formula to known Semitic morphology and demonstrates the power of computational linguistics.

His paper proposes readings for 37 of the script’s 102 signs, including all 13 unique to Linear A, and resolves three previously unknown Linear B values. Di Mino also assembled a 383‑entry English lexicon and a nine‑page draft titled *Ya Diktu*, poised for peer review. The methodology suggests AI could accelerate decipherment of other lost scripts. He credits the systematic symbol categorization and hypothesis testing to the Claude‑driven pipeline, arguing that similar AI‑assisted frameworks could unlock Egyptian hieratic or the Indus script for future scholarly collaborations.