The Voynich Manuscript: Understanding the Untranslatable with Lisa Fagin Davis

[eggyk]

English is not my native language. I have been criticized for repetitive phrasing when I wrote in my own English and also now that I use language tools. In both cases the content is mine. Every source is verifiable. Every quote is checkable. If any of my points is factually wrong, I would welcome a correction. The style of the prose doesn't change whether Bennett 1976 exists, whether Bowern wrote 'does not prove,' or whether Reddy & Knight found no vowel-consonant separation.

You are not just using the tool to formulate the english language or to just using it to translate your words, but instead to generate the entire structure and flow of each post (at least). I'm highly doubtful that you are actually writing a draft post in your native language and then asking the LLM to translate it. 

But fundamentally, the issue is that it changes whether we know that we are actually speaking to you. If the prose looks identical to what would be seen if you simply prompted "write me an in-depth post laying out my *insert argument here*, using the following *insert sources here*" how can we know that it is genuinely your content? Whether that content is true or not is an entirely seperate issue; it's the principle of knowing that we are actually communicating with eachother in a forum environment that matters here. It's noteworthy that the post that I am replying to/quoting actually feels like a post written by the real you. 

Plenty of the AI assisted theories that arose here before the great moderator firewall were -on the face of it- reasonable sounding theories with interesting ideas. However, the amount of time needed to check them for mistakes was so large compared to the time taken to generate them that it became a huge issue. 

However, I will say that it's clear that you are checking these outputs for mistakes, and are likely supplementing many details into them to ensure they are correct. But again, I can't know that for sure.