this post was submitted on 01 Oct 2025
106 points (100.0% liked)

Data Is Beautiful

8898 readers
22 users here now

A place to share and discuss data visualizations. #dataviz

founded 4 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
[–] gasgiant@lemmy.ml 5 points 1 day ago

My guess would be that they were manually written by individuals then certain words did overlap and become more common. Which would be normal for the language of a system.

However now the speeches are being written by the essentially the same author. Which works by looking for patterns and repeating them. So those common words become even more common. Which the model then picks up and makes sure to include more. Repeat until no diversity exists.

That's the actual future of AI.....