Boris Cherny claims to crack the undeciphered 3,500-year-old Linear A script using Anthropic's Claude Code

@bcherny This is incredible

@khoomeik @bcherny Hmmm extremely suspicious that everyone interested in deciphering Linear A with AI is from the mid hudson valley.

@bcherny this is obviously an extraordinary claim; many researchers have tried to crack Linear A for decades and failed. however, if it is true, there are going to be a lot of people who are forced to re-evaluate their perception of AI.

@eigenrobot turns out it's all dick jokes

Doesn't seem correct to me.

It claims Linear A is an extinct Semitic language. Almost none of the pre-Greek substrate is Semitic, and the phonology of pre-Greek seems mismatched with Semitic to such a degree that I've never heard of anyone claiming a connection between the two. And this despite many wild theories and speculation about connections of pre-Greek to other languages.

@eigenrobot @AngularOcean I expect to see a lot more of this soon

Guessing we’ll get animal communication in the next 5-10 yrs for some species (dolphins? Whales?)

@bcherny This looks awesomely promising. Next: the Voynich manuscript.

@bcherny cc @_aidan_clark_

@liminal_warmth @eigenrobot @AngularOcean Have you seen the progress on whales? Its 90% done

this is happening over and over in the replies

i don't understand a word of it. its entrancing

@lila_hollis @seconds_0 @liminal_warmth @AngularOcean we're gonna have a lot of explaining to do

The issue isn't borrowing the alphabet, it's that we know some pre-Greek words/suffixes and they don't fit Semitic languages.

For example, the -ινθος suffix in words like λαβύρινθος (labyrinth) is pre-Greek and isn't Semitic.

Similarly θάλασσα/θάλαττα (thalassa/thalatta), meaning sea, is a pre-Greek word and also not Semitic.

@bcherny This project reminds me @natfriedman 🌋Vesuvius Challenge

https://scrollprize.org/

@appelbolt @eigenrobot Just because Linear A may have been from a Semitic language does not necessarily that Minoans were a Semitic people.

@eigenrobot https://en.wikipedia.org/wiki/Pre-Greek_substrate#Phonology

Etruscan has nothing in common with Greek, but they derived their alphabet from the Greek one.

Latin, likewise, adopted many letters from Greek and Etruscan, but it’s all different words and only roughly similar sounds.

Why couldn’t a people speaking a proto-Greek without an alphabet adopt one from a Semitic people that had somehow ended up on Crete and developed writing?

@bcherny A cooler way to use Claude Code:

With Fable 5.

@bcherny Seems very promising, far simpler methods and less compute used to crack Linear B back in 1952.

@MatthewTanous @lefineder @eigenrobot "the Minoans probably didn’t call it Knossos."

We have two third party sources for the Minoans calling it something similar to Knossos, Greeks (Ko-no-so) and Egyptians (Ku-nu-sha).

@appsicle_ @eigenrobot No, modern cryptography can't be broken by this kind of statistical analysis.