Hey @jrgong, I’m less of a fan of this idea, as it’s not really clear to me that GPTs are very good at summarising technical documentation. See for example:
Also, I’m going to re-write the custom wizard plugin documentation soon.
Totally with you there, AI tends to hallucinate. Also, it depends on what provider is used and if custom prompt additions are added. E.g. adding prompt like “Do not hallucinate” and others can usually lead to somewhat good results. Or a user can ask “are you sure, can you verify this again”.
On our end, we use it for our staff to come up with drafts for documentations which are then corrected manually.
In the end the any AI-powered bot should be labeled as experimental to manage users’ expectations.