Did an Amateur Just Crack the 120-Year-Old Linear A Mystery Using Claude Code?

Tom Di Mino, a self-taught engineer from the Hudson Valley, is making waves by claiming he has finally deciphered Linear A, the Bronze Age writing system that has stumped experts for over a century. To put the difficulty in perspective, this is a script used from 1800 to 1450 BCE, leaving us with a tiny, frustrating corpus of roughly 7,500 symbols—the equivalent of less than two pages of text. While Michael Ventris successfully cracked Linear B back in 1952, Linear A has remained a ghost because it shares symbols with its successor but keeps 13 unique characters that have never been decoded.

The hypothesis centers on Linear A being an extinct Semitic language, a distant ancestor to Hebrew and Aramaic. Di Mino utilized Claude Code not to solve the puzzle for him, but to build a robust system of Python scripts that could cross-reference the GORILA and SigLA databases. By running over 100,000 statistical simulations, the AI allowed him to sift through linguistic patterns that would have taken a human researcher an entire lifetime to map manually. He claims to have assigned phonetic values to 40 characters and compiled a dictionary of 408 terms.

The scientific community remains, predictably, in 'show me the data' mode. Critics are already pointing out that with such a small dataset, it is mathematically easy to 'force' a solution that looks correct but lacks structural integrity. While Di Mino has drafted a nine-page paper, it has yet to pass peer review or be published in a reputable journal. Ultimately, this is a classic case of 'move fast and break things' applied to ancient history; we are left watching a high-stakes standoff between modern computational brute force and the stubborn, silent limitations of archaeological records.

Source: Claude Code Minoan Linear A