@ellis remember to let us know (via a comment here) when you feel your article is ready for review.
Thank you.

Thanks everyone,

Please review https://fedoramagazine.org/wp-admin/post.php?post=42917&action=edit

@ellis I've done a review of your article and in general it looks pretty good. I did a small amount of "word smithing" but nothing major.

I did change the title of one section to "Hooking it in: Connecting the RAG database to the AI verbally" (note the addition of the word "verbally"). When I first read this I wasn't certain where it was going and I thought that might clarify it a bit. (?) In fact, that section seemed to be less integrated to the rest of the article in some manner. If you can elaborate a bit more there it might be clearer.

I also was not certain about the one block of text that appears as follows:

Relevant Fedora Documentation:

$CONTEXT

I see CONTEXT used in the script talk.sh later but it wasn't clear what that reference was about untii I went searching for the variable. Perhaps indicate that the talk.sh script is also included and point out how the variable is loaded there ??

@rlengland Thanks very much for looking it over, it does bring to light how I didn't explain a certain part so well. I made some changes.

I think what is a little confusing is that the results from the RAG database are pretty much a verbatim answer for the question we are asking the AI. But they still need to get run through the LLM. (does that make sense? You want to go from asking the LLM "You are a helper, answer this question, what is Foo?" to asking the LLM "You are a helper, the documents say a Foo is a kind of tropical bird, answer this question, what is Foo?")

Please let me know if I can make this better, any comments welcome,

Ellis

@ellis I have completed my review. I did a few edits to improve readability (nothing major) and I modified a few link destinations. I think it looks good.

I would suggest that you give it another review to verify the links and make certain my wording changes have not modified your intended meaning.

One thing I would caution you about. Even though you provided a good "attribution" section, there are some who will have a "knee jerk" reaction simply because this article addresses AI. Be prepared for them. Be ready to address the technical side of the article.

Since all comments on articles are moderated, If they get too personal or obnoxious and don't directly address the content of the article we will not allow them to be viewed. But the tone of some articles may still be hard to deal with. Let us hope this will not be an issue. If you question a comment or possible response, please feel free to contact us on this ticket.

Let us know the results of your final review and we will schedule publication.

Thank you for your contribution.

@rlengland Thanks very much

I re-edited only the part about "skipping ahead". I tried to pick better wording to make it more clear. The idea is not to redirect them to my blog but just to send them further down the page and skip over the "create a database" section. If that looks good to you then everything looks good to me and we are ready to go!

I appreciate your warning and concern, thank you. I'll be fine with any comments, even very spicy and/or off base ones. One of the reasons I want to publish this is to connect with people who are not yet on board with AI.

@ellis I have set your article for publication at 08:00 UTC 6 February.

Thank you again for adding your article to the Fedora Magazine.