I wonder if there's anyone here experimenting with LLMs?
How is it working for you, and I wonder how you're planning to use it, or is it just experimentation?
tangor
4:13 am on May 27, 2023 (gmt 0)
At present LLM is a bit on the fuzzy side, paraphrasing, and not principally anything other than one more tool in the tool belt.
Not relying on it.
LLM is only as large as the current model and most times is "behind the times" and not up to date. Even in generic mode it is prone to error and requires fact checking by humans to be of any use at all.
That said, speaking of humans, they tend to take the route of least resistance and will tend to use these results AS IF TRUE and we'll have a potential for a lot of garbage results out there.
Remains an "experiment", not quite useful at the moment.
