Turns out that LLM summaries are actually useful.
-
Turns out that LLM summaries are actually useful.
Not for *summarizing* text -- they're horrible for that. They're weighted statistical models and by their very nature they'll drop the least common or most unusual bits of things. Y'know, the parts of a message that are actually important.
No, where they're great is as a writing check. If an LLM summary of your work is accurate then indicates what you wrote doesn't really have much interesting information in it and maybe you should try harder.
-
They're weighted statistical models and by their very nature they'll drop the least common or most unusual bits of things
yes!If an LLM summary of your work is accurate then indicates what you wrote doesn't really have much interesting information in it and maybe you should try harder.
no!
Just because something is statistically plausible does not mean it contains nothing of value that needs to be said. There's no correlation between text and meaning in an LLM, and there is absolutely no correlation between text and value or worth in an LLM, either.
An LLM may reproduce a statement that is true because it appears in many places; it does not mean it's inappropriate or without value to communicate! It has no ability to assess the point of what you're saying, and even less to evaluate the context in which it's said. -
Maho Pacheco 🦝🍻replied to Dan Sugalski last edited by [email protected]
@wordshaper I could not agree more. I am just experimenting with converting from a blog post to a thread using LLMS.
Check the thread: https://hachyderm.io/@mapache/113359910498119201
And the blog post: https://maho.dev/2024/09/my-coffee-history/
It got rid of the parts I mostly like. A shame.
-
@mapache @wordshaper did it get the points and just miss parts you like, or did it miss major highlights of the post?
-
Maho Pacheco 🦝🍻replied to Shiri Bailem last edited by [email protected]
@wordshaper it actually did a good job on summarizing and get the points, it did not miss the big parts. But as @shiri mentions, the points that make the post special (nuanced, human) got "washed".
For example, the blog post talks about coffee as a character: "Coffee in my childhood was always a supporting character in the main story."
Also, some of my fav phrases got removed (e.g. "as we say in Latin America, to be young and not revolutionary is almost a biological contradiction.")
