[XCSme]

Doesn't this make the "AI" even less creative and more like full-text-search instead? What makes some data a "fact"? If everything is written in the training data, in the end, won't everything be treated like a fact? So the LLM will have 100% accuracy and 0% creativity.

[qeternity]

Some people don’t want their language model to have any creativity.

[AlexCoventry]

"Ten times less" is a common English usage with a clear meaning.

[esafak]

Creativity is clearly not the goal here. Machine learning models are trained to be robust to errors in the training data.

[Brananarchy]

Most advertised commercial uses of "AI" are glorified search. This seems like an improvement in that space?

[AbstractH24]

Whats wrong with that?

If "AI" simply allowed people to search data without having to structure it then apply very rigid search terms, that would be in a win in my book.

[XCSme]

Nothing wrong, only that AI seems to be transforming into the next search engine, as opposed to AGI.

[AbstractH24]

Yeah AGI is way off, if ever.

Question is do the public and investors realize that.

[batch12]

Sounds like compression to me

[XCSme]

"Compressing" 500MB of data into a 70GB model.

[tzekid]

Isn't it more like "compressing" interactions with and patterns of 500MB of data?

Like, it's not only text search but also has a degree of generalization (e.g. in labelling new documents)