I have been experimenting more with various chat bots available on the Internet. Inkling, a new LLM from the American company Thinking Machines, is a higher-end LLM (although not really frontier). I tried having it rewrite a few Rust functions.
It reasoned through the statements and carefully designed faster versions. I liked how it gave me several rewritten versions: ones that use external libraries, ones that use just the standard library.
• I ended up applying 2 or 3 of the rewritten functions to my code base.
• Inkling wasn't particularly fast, but it was fairly cheap.
• Inkling had good responses—it is a nearly 1 trillion parameter model, and is open weights, so potentially could be run locally.
Basically my prompt is "could this Rust function be written in a faster way" followed by a paste of the function text. Each query takes like a cent and a couple minutes. I ended up with a few nice improvements. Inkling worked well, as did Muse Spark—I can't really say which is the better model.