A developer shared their experience using ChatGPT and Gemini to troubleshoot issues with gsoap and Onvif protocols. They found that the LLMs often provided suggestions that led to more errors, requiring numerous attempts and careful code reversion to achieve a working solution. This led the developer to generally reject pull requests generated by AI tools due to their tendency to introduce unrelated changes. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT Highlights potential pitfalls of relying on AI for code generation and debugging, suggesting careful human oversight is still crucial.
RANK_REASON Developer's personal experience and opinion on using AI tools for coding.