Oh lord, JCL. Brings back memories. I once had a summer job at IBM in which I ended up being the expert to which people would bring their JCL bugs, even though that wasn't my job in the least. :~)
ChatGPT is telling you about the really easy case in which a JS error just has one meaning. We should certainly reword those, and as I said we've done that in some cases. I think stack overflow may be more complicated.
Although, @jens, now that I think about it harder, maybe I'm misunderstanding these cases; maybe recursion base case errors are manifested in our heap-allocated Snap! process stack, rather than JS's stack? In which case I don't understand why we would ever get a JS stack overflow unless there's a bug in a primitive. I guess I should reread lists.js yet again...