In September I published what a language model found across 400 AI-built repositories. The most useful criticism was also the most obvious: a model's judgement is not reproducible, and you cannot audit it. Run it again and you might get a different answer.
Source: [Dev.to](https://dev.to/vibesolutions/i-re-scanned-393-ai-built-repos-without-using-ai-1-in-8-shipped-a-critical-flaw-2g5e)