Without getting into "good" vs. "bad", let's evaluate AI tools on a fairly simple metric - repeatability.
When I use a tool I expect to get the same outputs for a given set of inputs. Good compilers, assemblers, table saws, etc. don't change their output every month or two. I have a sinking suspicion that even if you document all of the prompts and nudges to the AI tools, you wouldn't get the same results each time. Based on that metric, you really can't compare an LLM to compilers. Unless you expect your compilers to do random things each time you feed them the same source code.
_______________________________________________ Freedos-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/freedos-devel
