Post #1791010

"AI is often incorrect!" is not a useful framing. A useful framing is a confusion matrix, plus the ability to exclude false positives (e.g., with constraints). The same as any linter. Then you can judge the usefulness.