Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Integrate ChatGPT (or similar) to AndBible #3302

Open
tuomas2 opened this issue Aug 15, 2024 · 6 comments
Open

Integrate ChatGPT (or similar) to AndBible #3302

tuomas2 opened this issue Aug 15, 2024 · 6 comments

Comments

@tuomas2
Copy link
Contributor

tuomas2 commented Aug 15, 2024

It is possible to instruct ChatGPT to make andbible-understandable links from all bible references. Simple chat window would be already useful for this purpose. Perhaps with some instructions to AI to make answers more useful for bible study purpose.

Negative side: user needs to obtain OpenAI API key, which is costly and tricky as they only allow this through contacting sales.

@tuomas2
Copy link
Contributor Author

tuomas2 commented Aug 15, 2024

Nicodemus.ai is something we could consider too.

@GWev
Copy link

GWev commented Aug 15, 2024

It would be a reason to leave the app or not update anymore. Prefer to study by my own, with G's/help. Data protecting. Everyone is free to use such technology independently. Other issues are more important, such as free open exporting of notes and favorites ( export bookmarks & notes in txt / csv / ods.. #3235 ). Thank you

@tuomas2
Copy link
Contributor Author

tuomas2 commented Aug 15, 2024

If implemented, it would be opt-in feature and not enabled by default. But good to know your reaction.

@castek
Copy link

castek commented Aug 16, 2024

Tool for non English speaking readers

There are plenty of biblical studies/thoughts in English on internet, but other languages might have poor amount of them. How to make English texts accessible for non English speakers? Either by some translator that will react on non English question by automatic translating the question to English, automatic querying the web in English, and automatic translating results back to non English language. Or by some web language model (eg. like ChatGPT) that knows almost all the topics discussed on web and can communicate many (or all?) languages used on web. Integrating language model to AndBible and asking the question in any language can make all most discussed biblical topics in any language with AndBible-undestandable links to the scripture accessible to readers of all languages.

Contextual search

Sometimes I remember the idea of some verse and not very successful in text search through Bible. Asking "What biblical verse is saying something like your ... is new every morning?" to some language model can even omit all web sources and focus just on some or more Bible translations and find what I'm looking for and show the verses directly in AndBible.
Other queries: "What biblical text is connecting parts of body with gifts?" I assume Romans 12:4-8 as a result, but at least Romans 12:4-6 since verse 4 is talking about body and verse 6 about gifts.
"Find the closest similar biblical text as in Matthew 6,9-13?" I assume Luke 11,2-4 as a result.

Translation variance

In translation comparison view AndBible can order different translations of the same verse using bible language model (or web language model) with according to translation similarities. Eg. for Romans 9:5 it will show NIV with NASB translation at one end of translation comparison list and NRSV with RSV at the other end, because NIV and NASB are similar but different from NRSV and RSV.

User may be able to switch between just biblical language model (built on different Bible translations without web) and web language model (knowing both the Bible translations as well as web).

User can search ChatGPT & similar without AndBible, but of course when language model is integrated in the AndBible, user needn't to switch between apps and can do contextual search of displayed text, or display the result directly in AndBible and its native views. Personally I'm not content with czech (my primary language) cross references, they cannot be used very well, so I'm substituting it with different instruments.

@GWev
Copy link

GWev commented Aug 16, 2024

I've read your posts with attention. My first comment may seem a bit radical. I will try to explain:
@castek
The point you make is about quantity of information and rapidity.
As we learn in times of data multiplication, more is not often better. The more possibilities exist, the more information is available, the more it needs also wisdom in order to use it in a proper way.
What the bible reader needs is the right attitude, G's leading by his spirit, persistence and so on.
I've encountered the mentioned issues several times (makes me even "angry" sometimes when I search for the right words, but no result because of translation, right order or spelling issues - even the smallest), but I kept searching with prayer, trusting his timing and leading, very often with success.
Again, I understand each of your points, and ideas for tools which can be very helpful. But it takes time to process. Just adding a flood of information from a non-trustable source (which is Chat GPT, the result at least can be biased, we know already many examples) is not the solution. We cross a line here, that's my conviction. Maybe for some it is kind of "philosophical" question.
If possible, a better text-based search could be helpful, for sure. But I find it also rewarding when I have to "meditate" on G's word i.o. to find the right passage, the right words, thinking about the meaning and possible words to use.
Surely a question to meditate about!

@castek
Copy link

castek commented Aug 16, 2024

Thank you, I understand your opinion.
I'm also aware that language models can answer misleadingly and I as well as all language model users have to check the correctness of models' answers all the time.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
Status: Prio 2
Development

No branches or pull requests

3 participants