news.volyx.in

AI Engineer Claims to Have Cracked Linear A (aiclambake.com)

448 points by Kosturdistan · 70 days ago · 182 comments on HN

Article summary

Tom Di Mino, an AI engineer and amateur linguist, claims to have deciphered Linear A, a Bronze-age Minoan writing system that has eluded linguistics experts for over a century. His claims are being reviewed by experts at Rutgers and Cambridge. Di Mino used Claude Code to analyze and organize the Linear A corpus, and his insights suggest that Linear A maps to an extinct Semitic language. He has proposed readings for 40 of the script's signs and translated 408 Linear A terms into English.

Main themes

  • Linear A decipherment
  • Linguistics and language analysis
  • AI-assisted research
  • Semitic languages
  • Ancient writing systems

What commenters say

  • The claim of deciphering Linear A is being taken seriously due to the credibility of the approach and the results, but some commenters are skeptical and want to see more evidence.
  • The use of Claude Code in the research is seen as a valuable tool, but some argue that relying on a stochastic system like an LLM can make reproducibility difficult.
  • Some commenters believe that the lack of a publicly available write-up or preprint makes it difficult to evaluate the claim, while others understand the need for peer review before publication.
  • There are differing opinions on the potential for the approach used to decipher Linear A to be applied to other undecipherable languages, with some seeing it as a promising development and others being more cautious.
  • Some commenters are concerned that the blog post announcing the claim is not a credible source and that more skepticism is warranted given the history of amateur claims in the field.
  • The importance of rigorous testing and verification is emphasized, with some commenters noting that even if the claim is true, it needs to be subjected to scrutiny and peer review to be accepted.
  • There is a debate about the role of stochastic systems in research, with some arguing that they can be useful tools, but others highlighting the potential limitations and need for careful evaluation.