Chat GPT / Open AI Plug-Ins
I have been using Chat 4 for a few days, and I can see a great potential if I can find the right plug-in. So far this is how I have been using it.
I keep a journal of my personal commentary in a word document. My journal commentary is strongly rooted in my Logos collection of many different sources, and I copy and paste whatever I site into my word document for easy reference. So far I have gone through several books of the bible and am only a few verses from finishing Revelations.
My personal method is this. I site a scripture. Then study all kinds of sources through Logos or print books that I have in my library. Once I find a few intriguing things I write about 3 paragraphs in a stream-of-consciousness sort of journal entry.
Where Chat-4 comes in is this. I uploaded several studies I finished including my journal and the referenced material. Then I ask Chat a bunch of questions such as: What is the common theme? What is the subtext? I go through as many textural questions and analysis as I can think of.
I had to ask it to refrain from it's own opinion about things, and just stick to my journal entries and ancient near-eastern ideas, because it kept interjecting things about environmentalism and earth stewardship and whatnot. But once I addressed it and told it not to do that it seemed to comply. I think it's just in how you phrase it. If you have trouble let me know and i can look up my exact chatlog and post it.
Through that I was able to upload several sections of commentary and other findings and re-evaluate my own journal entries. I asked it to look for logical fallacies and many other things. I even had it make a modern narrative using the story from Rocky but based upon my commentary of Proverbs 16.
But here is my dilemma. It will only handle so much text that I input into Chat-4, I think it goes up to 64,000 tokens (which I think are a similar idea to words--not exactly sure). My Revelation journal commentary is already triple that size. So I think with the use of the right plug-in I would be able to upload my whole commentary, as well as many other resources and could do a lot more with it.
I know I said a lot here, but I'm excited to see where this goes and if it or something like it can be integrated within Logos it could greatly empower our research and study. I know for some it may make them lazy, but for me it just stokes the flames even more, and I think there are many others like me.
Thanks
Comments
-
" I know for some it may make them lazy, but for me it just stokes the flames even more, and I think there are many others like me. "
that same thing could be said for tools in Logos I guess. but like you, the explosion of technology fascinates me and make me excited. it won’t go away, so the Christian should think about how it might, could, or should be used.
reading the new Steve Jobs book now (one employees got and the ebook was free) and it amazes me how every few pages is something that makes me think 'wow, im so thankful for that' though he had no clue me and others would be using it in Christian ways decades later.
preachertony.com — appletech.tips — facebook.com/tonywalker23 — twitter.com/tonywalker23 — youtube.com/tonywalker23
0 -
I recommend everyone read the other thread on chatGPT where it cites sources that don't exist and makes up quotes, and their attribution.
It's an interesting technology. But it's just not there yet for serious research. I'm sure it (or something like it) will be fantastically powerful in our libraries one day. But not this day.L2 lvl4 (...) WORDsearch, all the way through L10,
0 -
This is true, but you can set the parameters of what information it uses and check the results. It’s always good to check sources Ai or Logos, print commentates, etc. Ideally, Logos would have their own version engineered for their software and our work here. And thanks for your suggestion to read the other posts. I will do that. Peace.
0 -
I am finding Google Bard more reliable than Chat GPT. It sometimes provides references and it quicker to say it is not trained to provide the answer - both features I appreciate. One still gets answers that can't be verified e.g. the Africana Bible available publicly is an expanded Old Testament. Google Bard gives me a canon for a New Testament but I can't confirm that it exists in the Africana project.
Orthodox Bishop Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."; Orthodox proverb: "We know where the Church is, we do not know where it is not."
0 -
Thanks, I've never tried that one. I'll check it out. I assume Logos will eventually either build or conscript their own, or else get some sort of plug-in.
0 -
MJ. Smith said:
I am finding Google Bard more reliable than Chat GPT. It sometimes provides references and it quicker to say it is not trained to provide the answer - both features I appreciate. One still gets answers that can't be verified e.g. the Africana Bible available publicly is an expanded Old Testament. Google Bard gives me a canon for a New Testament but I can't confirm that it exists in the Africana project.
"YMMV" (your mileage may vary) definitely applies to this comparison! Here's someone who had the opposite experience:
0 -
If Logos does not step up into the AI game, I fear new and better software will emerge.
0 -
If you set the parameters and check the results they both work fine. This should be standard AI or no.0
-
I would be extremely surprised if this is not in the works yet.0
-
I've been comparing responses in Google Bard to ChatGPT (since MJ has made me aware of Bard's existence). It seems like ChatGPT has better-sounding language but Bard is (slightly) more accurate in its responses. Although both give faulty information, so far Bard seems less egregiously wrong. Bard also gives three drafts of each response and you can pick which one you like the best, although once you ask the next question you lose the other drafts.
0 -
Louis Vigo said:
This should be standard AI or no.
Regardless of how similar the underlying AI algorithms are, their training was clearly optimized for quite different goals. BARD will often respond that it is a language processor, if you push it too hard for minute factual information.
Orthodox Bishop Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."; Orthodox proverb: "We know where the Church is, we do not know where it is not."
0 -
I tired Bard also, since he made me aware of it. I set the parameters by telling them both to only use the information I give them, and I use them as a sort of textural analysis tool. I coped and pasted some pretty hard to understand commentary sections and it broke it down where I could understand it. Then I asked it some questions. I'll copy the chat log and post them both.
0 -
Here are the 2 chat logs from Bard and GPT-4. I asked both the same questions. And you can see that by my use, I am using it with Logos, not as a replacement for Logos, but as a kind of textural analysis. That way I can set the parameters on where it draws it's knowledge from and I can check it results since I gave it the info. I'm not a professional scholar so I'm sure one of the big guys would make way better use than I have.
0 -
Louis Vigo said:
Here are the 2 chat logs from Bard and GPT-4. I asked both the same questions. And you can see that by my use, I am using it with Logos, not as a replacement for Logos, but as a kind of textural analysis. That way I can set the parameters on where it draws it's knowledge from and I can check it results since I gave it the info. I'm not a professional scholar so I'm sure one of the big guys would make way better use than I have.
Good stuff.
Hey if you are interested can you contact me for more work on this.
I am also developing these same ideas as you.
Remind me that I know you from the forums.
Peace
0 -
Wayne levi Price said:
If Logos does not step up into the AI game, I fear new and better software will emerge.
As one who has been a long-time observer of Bible software companies, I would be stunned if any Bible software company took a plunge into using AI tech in a meaningful way. Sadly, Bible software as a sector of software industry is contracting and the publishers have such a stranglehold on their content that you have to be a big player to get anywhere with them. So a small upstart won't likely be able to do this and the big guys are not forward-thinking enough.
Dr. Kevin Purcell, Director of Missions
Brushy Mountain Baptist Association0 -
abondservant said:
I recommend everyone read the other thread on chatGPT where it cites sources that don't exist and makes up quotes, and their attribution.
It's an interesting technology. But it's just not there yet for serious research. I'm sure it (or something like it) will be fantastically powerful in our libraries one day. But not this day.I can't think of anything more antithetical to the God who personally communicates and instructs His people through the objective truth of His word, than artificial intelligence. Oh, it will start small and innocent enough, but eventually it will become a false authority for many. Jesus said, "I am ... the Truth."
Instead of Artificial Intelligence, I prefer to continue to rely on Divine Intelligence instructing my Natural Dullness (Ps 32:8, John 16:13a)
0 -
JRS said:
I can't think of anything more antithetical to the God who personally communicates and instructs His people through the objective truth of His word, than artificial intelligence. Oh, it will start small and innocent enough, but eventually it will become a false authority for many. Jesus said, "I am ... the Truth."
Agree. Not with the doctrinal statement per se. It's just that computers don't make good witnesses to the Truth. More like witnesses to the opinions.
But then, I think seminary folks should avoid dataset-searches (and RI's). Get comfortable learning, instead of watching.
0 -
Cogitating under our fig tree is a lost art in our Google/Wiki/AI world. John 1:45-50
Instead of Artificial Intelligence, I prefer to continue to rely on Divine Intelligence instructing my Natural Dullness (Ps 32:8, John 16:13a)
0 -
JRS said:
I can't think of anything more antithetical to the God who personally communicates and instructs His people through the objective truth of His word, than artificial intelligence.
Ignoring the theology implied, I remember when I was 9-10 being fascinated by the charts in Halley. All the "esoteric" knowledge neatly arranged for easy digestion. Those charts got me into Bible study. I think AI has some of same potential ... if it ever get to the same reliability level as Halley ... reliable if you buy into his perspective. At the moment the chatbots are too reliant on how you ask the questions and too willing to provide pitchfork material (think mucking out a stall).
Orthodox Bishop Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."; Orthodox proverb: "We know where the Church is, we do not know where it is not."
0 -
JRS said:
I can't think of anything more antithetical to the God who personally communicates and instructs His people through the objective truth of His word, than artificial intelligence.
Ignoring the theology implied, I remember when I was 9-10 being fascinated by the charts in Halley. All the "esoteric" knowledge neatly arranged for easy digestion. Those charts got me into Bible study. I think AI has some of same potential ... if it ever get to the same reliability level as Halley ... reliable if you buy into his perspective. At the moment the chatbots are too reliant on how you ask the questions and too willing to provide pitchfork material (think mucking out a stall).
Orthodox Bishop Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."; Orthodox proverb: "We know where the Church is, we do not know where it is not."
0 -
I'm with you on this, I see this as being useful for queries that can be asked in natural language, then search through my library and report back. The ability to have it produce a sermon or bible study (or your traditions equivalent) wouldn't be desirable for me. But the AI's have to stop being creative with the truth before they will be useful. If I can't depend upon the results, then what would it benefit me?MJ. Smith said:JRS said:I can't think of anything more antithetical to the God who personally communicates and instructs His people through the objective truth of His word, than artificial intelligence.
Ignoring the theology implied, I remember when I was 9-10 being fascinated by the charts in Halley. All the "esoteric" knowledge neatly arranged for easy digestion. Those charts got me into Bible study. I think AI has some of same potential ... if it ever get to the same reliability level as Halley ... reliable if you buy into his perspective. At the moment the chatbots are too reliant on how you ask the questions and too willing to provide pitchfork material (think mucking out a stall).
I see in the not so distant future, an AI trained on my library's index, that I can ask "what is *author names* position on *bible verse*". Or something like that and get an answer based on the contents of my digital and perhaps print version of books that Logos owns rights to - Complete with clickable links to the context.
And yes, they should be able to read the books that are simply scans of books too. OCR would be important in such a scheme..
L2 lvl4 (...) WORDsearch, all the way through L10,
0 -
Yes, I agree that there would be many downsides if mishandled. but truth be told pastors have been buying sermons for many many years, so an AI writing a sermon would be nothing new. I think people are more afraid of the artificial life part. There are many AI's that are already a part of our lives and we don't even know it. Insurance companies, engineering, marketing, technology manufacturing firms use them already. Much of the technology in your house uses a form of AI, it just doesn't have a language interface. And maybe that is what scares people, because it seems lifelike. And I get it. It fits the profile of the Beast in Revelation.
But apart from speculation, I'm more interested in what you said, a language based interface to all the datasets I have in Logos, a way to search, compare, contrast, and connect ideas. Having a tool that can help you edit, look for logical fallacies, repeating themes, phrases, tone, etc.
I wrote some short stories and got some valuable feedback by asking it where a reader might get lost or bogged down. Also, I had it simplify a section of a technical commentary, then compare that to another commentary article. Yes, you have to check it's work. Don't trust it. But that should be a given that were sort of modern Bereans.
0 -
ChatGPT is a wonderful tool and will be used for good and evil. It is almost certainly an inevitable part of our future and an imperative part of Logos' Future. Logos should not wait for the next iteration of Logos to incorporate it. Other software companies are already doing amazing things with it (thinking the medical field and Adobe).
I use AI extensively for all kinds of Bible and other research Bible to provide an amazing amount of structured data in a hurry and then use Logos to validate what parts of the data I can trust. So far, I would anecdotally rate AI returns at 100% helpful and 90% reliable, so verification is imperative.0 -
I don't think you guys need to get excited over negatives. FL already crossed the river with translation. No, doubt 'AI' is next. I think the only question is whether before L11. What, a year and a half or so.
0 -
GaoLu said:
So far, I would anecdotally rate AI returns at 100% helpful and 90% reliable, so verification is imperative.
I would rate them more:
- 50% BS - filler telling you what you already had to know to ask the question
- 80% answers are helpful
- 60% answers are reliable
- 80% of answers are incomplete
An example of the problem, I will ask it what the best selling Swedish Bible is; it gives me an answer I assume is accurate. I ask it whether that Bible includes the apocrypha. It will tell me it does not. I point out that many Swedish Bibles include it. It responds oh yeah, it appears between the Old and New Testament but isn't considered canonical by all churches ... and gives me a paragraph or two about the canonical status of the deuterocanonicals. I fully expect the chat boxes to become standard use quite quickly. My concern is all the hype around them and the lack of discernment over the amount of detail they hide and the biases they exhibit e.g. the assumption that the 66 book Bible is standard despite it being the minority world wide.
Orthodox Bishop Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."; Orthodox proverb: "We know where the Church is, we do not know where it is not."
0 -
[Y]
0 -
I suppose its fine when its my minority view that is agreed with, but not fine when it is someone elses.MJ. Smith said:GaoLu said:So far, I would anecdotally rate AI returns at 100% helpful and 90% reliable, so verification is imperative.
I would rate them more:
- 50% BS - filler telling you what you already had to know to ask the question
- 80% answers are helpful
- 60% answers are reliable
- 80% of answers are incomplete
An example of the problem, I will ask it what the best selling Swedish Bible is; it gives me an answer I assume is accurate. I ask it whether that Bible includes the apocrypha. It will tell me it does not. I point out that many Swedish Bibles include it. It responds oh yeah, it appears between the Old and New Testament but isn't considered canonical by all churches ... and gives me a paragraph or two about the canonical status of the deuterocanonicals. I fully expect the chat boxes to become standard use quite quickly. My concern is all the hype around them and the lack of discernment over the amount of detail they hide and the biases they exhibit e.g. the assumption that the 66 book Bible is standard despite it being the minority world wide.
Has any one asked it questions to try and nail down some of the other biases it has beyond a protestant canon? Can we nail it down to a particular tradition or denomination stream?L2 lvl4 (...) WORDsearch, all the way through L10,
0 -
As long as we can quantify its bias, then our expectations can be better managed.
In asking it some more questions it seems as though the decision tree regarding apocrypha questions is as follows:
Is the translation protestant?
If yes If no
No Apocrypha Apocrypha included.
Its not actually checking the translation it self, its making assumptions based on probability.
The church of Sweden is lutheran, and chat gpt considers lutherans protestant, and assumes they will always use the protestant canon.L2 lvl4 (...) WORDsearch, all the way through L10,
0
