# Can LLMs (large language models) be used to decipher ancient languages? **URL:** \ **Category:** In My Humble Opinion\ **Tags:** ai\ **Created:** [July 8, 2025, 11:54am UTC](https://boards.straightdope.com/t/can-llms-large-language-models-be-used-to-decipher-ancient-languages/1020628 "2025-07-08T11:54:35Z")\ **Posts on this page:** 1\ **Showing post:** 38
**Author:** ![Stranger\_On\_A\_Train](https://avatars.discourse-cdn.com/v4/letter/s/13edae/32.png) [@Stranger\_On\_A\_Train](https://boards.straightdope.com/u/Stranger_On_A_Train)\ **Post date:** [July 9, 2025, 4:04am UTC](https://boards.straightdope.com/t/can-llms-large-language-models-be-used-to-decipher-ancient-languages/1020628/38 "2025-07-09T04:04:04Z")
> [@leahcim](#): > > I claim that the machine translator _is_ doing what the natively bilingual person does. It is reading the input, forming a semantic model of what is being described, and then describing the semantic model in the other language. The internal state of the LLM is a representation of the situation being described by the input, not just the input itself. LLMs do not have a ‘strong’ (conceptual) semantic model. They have a ‘weak’ semantic model basic upon a large array of textual associations but as demonstrated by @Pasta in [this post](https://boards.straightdope.com/t/future-of-general-questions-in-the-ai-age/1020443/65) and [this post](https://boards.straightdope.com/t/future-of-general-questions-in-the-ai-age/1020443/66) it makes basic errors and often just fills a response with ‘word salad’ when a subject is too complicated or highly technical to be represented by high frequency word and sentence associations. Stranger --- _[View the full topic](https://boards.straightdope.com/t/can-llms-large-language-models-be-used-to-decipher-ancient-languages/1020628)._