The more I use ChatGPT, the more I tend to learn about its intricacies and nuances. One of my latest discoveries is the “Are you sure?” hack:
![](https://i0.wp.com/judepereira.com/blog/wp-content/uploads/Screenshot-2024-06-30-at-09.21.21-1.png?resize=946%2C1024&ssl=1)
Yet another example:
![](https://i0.wp.com/judepereira.com/blog/wp-content/uploads/Screenshot-2024-06-30-at-09.30.21-1.png?resize=733%2C1024&ssl=1)
Both chats were with 4o.
What’s the take away? Make ChatGPT second guess itself, until it arrives at a concrete answer. Lastly, always double check critical information :)
Leave a Reply