Angry Russians displaced after Ukraine crossed the border and invaded the Kursk region last week have vented their frustrations online to President Vladimir Putin.
The criticisms represent an unusually public show of defiance in a country where any cracks at the leader or military can draw harsh punishments.
Understandable if the comment is entirely LMM generated, but to imply I should post a disclaimer every time one is used for summarizing content is a bit of a reach IMHO.
LMMs are a tool to be used, like anything else.
Summarising is one specific use of llms that doesn’t actually work.
Strange how it seems to work for me…
I know precisely what I want to say, I’m just asking for the information to be condensed into a concise 1-2 sentence statement.
There’s a big difference in asking it to generate something wholesale vs. feeding it information and asking for that information to be summarized in a clean and easy to understand format.
I would argue that is the best way to use LLMs; it’s basically acting as an editor.