Instead of spinners, you should use progress bars. When the endpoint of a process is unknown, you should just have a ticker that displays an increasing number. In codex, for example, just show me the number of tokens read or generated so far. It gives me feedback that the process hasn’t frozen, and it provides useful information. What an easy and cheap solution!

I’m having trouble passing your comment. Are you saying I’m more stupid than a toddler, or that the marketers are?

LLM stands for “large language model” in other words, it is a big neural machine. Saying they don’t use LLMs is like saying the ocean isn’t blue, it’s azure.

I’m not that well versed in rust, is this literally just wrapping the C library in rust, and calling it a rust implementation of it?

I think there are programming paradigms that haven’t even been discovered yet. Rock music was “discovered” centuries after baroque music, and composers back then would have thought baroque music is the ultimate musical genre, which could not have been improved upon.

I’ve been working on a programming language recently that sort of programs itself. It’s hard to explain how it works, since I approached it from a mathematical direction, but for a domain and range defined with a set of correct input-output pairs, there are an infinite number of functions which correctly maps one input to its corresponding output. If we define f_1 as all functions which correctly map the first IO pair, f_2 as functions which correctly map the second IO pair… Etc etc until we reach f_N, then the set intersection of all f_1…f_N is the set off all functions which satisfy our requirements. From this set of known correct functions, we can sort them by speed or some other trait.

But the devil’s in the details there. I won’t go into how I get f_1…f_N, but I’m having fun thinking about it & designing it. It may even be useful some day.

deleted by creator

deleted by creator