AI is not very good at reasoning from scratch, it depends on how well you craft 
the prompts. If you want to get better results, you need to spend more time on 
designing the prompts and defining the task. For example, I used the whole 
GitHub [issue](https://github.com/planetis-m/naylib/issues/84) with the 
discussion as a prompt and it gave me some useful hints. This helped me to 
discover a buffer-overflow bug.

Reply via email to