Date: October 7th, 2026 7:52 PM Author: sickened fishy stag film
progress can be made for sure. doubt it can be fully deciphered without discovery of an additional text or discovery of an unknown connection to a language we already have deciphered. if you attack it try that angle. try letting astra to a search comparing it to whatever other languages we have that have a good chance of portions being related to it
Date: October 7th, 2026 7:57 PM Author: out-of-control point mad cow disease
I wonder how many texts in Linear A and B and how much Scholarship about them is already in the corpus AI draws from and analyzes, and to what extent you'd need to do the leg work of going out and getting it.
Date: October 7th, 2026 7:59 PM Author: sickened fishy stag film
huge amounts are already in its weights, it has a lot of things "memorized", but also it will find everything additional for you. if you use ultra especially it is easy to just delegate subagents to go do searches for the most relevant texts and update the other subagents as it finds stuff.
Date: October 8th, 2026 3:06 PM Author: sickened fishy stag film
I can take an initial crack at it if you guys want. we need to all make anon github accounts and a central xo organization for any ideas poasters have they might want to pursue. have considered this before im just lazy. i dont feel like changing settings in my terminal to not doxx myself when it puts the name of your machine and email when you make commits. im sure it takes like two seconds to do a set up where you can use multiple identities but I'm lazy about stuff like that.
I'll try having it analyze against Hurrian, anatolian languages, ancient semitic languages and Egyptian.