9 Comments
User's avatar
Daniel Tenreiro's avatar

Shako this is awesome.

shako's avatar

❤️ thanks man :)

PapayaSF's avatar

I wish I had a clever or insightful comment beyond “Wow.”

shako's avatar

how kind!

Ned Twigg's avatar

I can’t wait to try this out! It relies entirely on the model’s internal sense of time, correct?

shako's avatar

thank you! yes the data i generated from this is entirely from gemini 3s huge memory from all its params. as i mention in the essay it definitely is a poc and a proper solution needs to ground it in a corpus of some kind

Daniel Tenreiro's avatar

Shako this is awesome.

Brenton Milne's avatar

Sick app

There's so many things I'd love to do with LLMs & historical texts

Like grade the predictions of every text, author, political ideology

DJ's avatar

Wow, this is really impressive. I like this approach to hazy categorization. It seems like it could be used for book ideas or even political campaigns.