I think LLMs are great at compression and information retrieval, but poor at reasoning. They seem to work well with popular languages like Python because they have been trained with a massive amount of real code. As demonstrated by several publications, on niche languages their performance is quite variable.
Comments
I think LLMs are great at compression and information retrieval, but poor at reasoning. They seem to work well with popular languages like Python because they have been trained with a massive amount of real code. As demonstrated by several publications, on niche languages their performance is quite variable.
I used to find it better to shortcut the AI by asking it to write python to do a task. Claude 4.6 seems to do this without prompting.
Edit: working on a lot of legacy code that needs boring refactoring, which Claude is great at.