I also agree with @Zoom. Using LLMs to generate documentation is fine assuming 
one proofreads, edits, and validates. Which means you are just using it as a 
fancy writing aid instead of relying on it being right. Especially given I have 
seen multiple people act like LLMs are gospel and ask questions like "Why did 
ChatGPT generate incorrect code?" or "Did this work in a previous version?". It 
is best to avoid publishing any generated content without this barrier as it is 
actively distributing misinformation.

If anything though this website Stefan has put together does demonstrate that 
it would be nice to expand the official FAQ and have it easily accessible on 
the website. For instance @DemoTomohiro's 
<https://internet-of-tomohiro.netlify.app/nim/faq.en.html> is quite fantastic 
as it's easily searchable for common issues.

Reply via email to