An Experiment with GPT4

[remod]

Thanks @my_key! I did not read the outcome in the original post, I'll do it right after I post this answer.
Here's the full reading of Yishi (which uses the Whilelm translation) of the response you got (5.6 > 9): https://chat.openai.com/share/9de692f4-575e-45b1-9b64-ee49d2b2d00c
summarized as:

Summary​

Yi (I Ching) ceases to speak when individuals close themselves off to its wisdom, particularly when they become untruthful, impatient, or unwilling to face and learn from unexpected challenges. Yi's guidance is continuous, found in both stillness and action, in waiting and in small, steady efforts. It's essential to remain open, honest, and patient to continue receiving Yi's insights. The hexagrams emphasize the importance of embracing challenges as opportunities for growth and maintaining a positive attitude during periods of waiting or gradual progress.

and this is the generated image:

The image inspired by your text has been created, capturing the essence of the Yi (I Ching) and its continuous guidance through truthfulness, patience, and learning from unexpected challenges. The scene depicts an ancient sage under a large tree, with gathering clouds above and three shadowy figures approaching, symbolizing the themes you described.

---

I've asked for an interpretation to Huiyi (which uses the Bradford Hatcher's translation) which gave me the following: https://chat.openai.com/share/0be93a5b-4a2d-4ecf-8dcd-01792652bdb7 summarized as:

Summary​

In the context of your question, "When does Yi stop speaking to you?", the I Ching speaks to the nature of guidance and wisdom as being ever-present, especially during times of anticipation and preparation (Hexagram 5). The changing line at 5.6 particularly suggests that guidance (Yi) appears even in challenging or unexpected situations, and engaging with this guidance leads to positive outcomes. The transition to Hexagram 9 further emphasizes that guidance is continuous, especially in periods where potential is high but results are not yet visible. Yi, symbolizing wisdom and guidance, does not cease to speak; rather, it is always available, particularly when nurtured and attended to in small, consistent ways. The key is to remain open, attentive, and engaged with the signs and lessons that come your way, even when they are subtle or gradual.

Here is the generated image:

and its explanation:

Here is an abstract representation of the I Ching's wisdom during times of anticipation and preparation, as interpreted from the provided text. This image captures the essence of guidance and wisdom in a visual form.

---

They seem to converge quite a bit (which is understandable as they are just the same only using different words for the text). It seems to me that, to further summarize both, they believe that the meaning of 5.6>9 is that the I Ching never stops talking to you but one may not be able to hear (or ready to follow) his advice.

I'm now curious to see how close (or far) this is from the interpretation given in the thread.

 

[my_key]

Hi remod
I'm blown away by the images. Being a visual person I can connect with them in ways that go beyond the words. I hope others can too. They provide a wholeness for the readings.
Possible titles that spring to mind for me "Sage sits with The One Tree" or "The Guardian of the Well"

 

[my_key]

Hi remod

There appears to be a dearth of takers for your offer to ask questions.

So I'm putting in a request for another go. I like the idea that readings by AI are devoid of bias and prejudice. They seem to carry a sense of purity about them. Again, I'm particularly intrigued to see the picture images generated on this.

This question is a bit older than the others but at the time I was well immersed in the path it led me down HERE

My intention when asking the question was not to impose any direction for the reading to prove entropy is the mechanism for change, or changing lines, more just a general 'So what do you have to say on this?' Although my convictions veer along the path that says there is a strong connection. This although unspoken could have influenced, in an unspoken way, the cast I produced, but I did endeavour to generate a neutral aura around the casting process.

“Entropy and Changing Lines?“

37: 3.4 <> 25

I hope you feel it is appropriate to follow through on this request.

Many Thanks.

 

[IrfanK]

Hi everyone! Long time, no see!

Hmm. Interesting experiment. And I tried to do exactly the same thing. I was learning how to code: a wonderful challenge for a 60 year old whose last encounter with maths was well over forty years ago. And naturally my thoughts turned to the idea of project with the Yi. I hacked away and drew up some fairly ugly code that could reliably print out the text from Wilhelm's translation. Feed in the numbers, it spits out the text. (It was actually quite a lot more challenging for me than it should have been).

Yeah and the next idea was to see whether ChatGPT could do a half-way convincing interpretation, applied to a real-life question. And the results looked good, at least if you didn't look that carefully. It all sounded plausible enough. I didn't like the style of the language much, so I experimented with that. At one stage, I even fed in a short passage from one of Hilary's blogs, and told it to imitate her language style. Yeah, it did look halfway plausible. I'm sure Hilary would read through it and say, I'd never say that! But if you didn't know too much, it sounded quite like her.

But in the end, I decided it was a silly experiment. It reminded me of a conversation I had with the personal assistant of one of Indonesia's oldest living "great" writers, one of the few people who lived through the struggle for Independence, witnessed it and took part in it, and could actually still talk about what he'd seen and done. Every week, he writes a column for a magazine, and people eagerly look forward to it. And the assistant told me that people had started producing fake essays, putatively written by the writer, and they made her feel a bit sick. They were sort of convincing, she said. They did catch some of his style.

But GM writes the way he speaks. Ninety-five percent of the time, he sounds like a grumpy old man, muttering nasty things to himself. And then every now and again, he'll say something so brilliant and sharp that it leaves you breathless, that you remember who he is and what he's seen and done.

The thing with the fake essays, she said, was that they got the 95% part, but they couldn't come out with those little flashes of brilliance. And that's what the ChatGPT readings felt like. The flat, mundane part of the reading. It doesn't create space for the flash of inspiration -- as Sylvia Plath said, that rare, random descent.

But yeah. ChatGPT is a confabulation machine. It's absolutely brilliant at producing writing that looks superficially plausible.

 

[Trojina]

Th absolutely brilliant at producing writing that looks superficially plausible.

But why would that be a good thing ? Why is it a good thing to fake real minds making real writing? That's a sort of rhetorical question because I don't think from you've said that you think a great deal of that.

I like the idea that readings by AI are devoid of bias and prejudice. They seem to carry a sense of purity about them. Again, I'm particularly intrigued to see the picture images generated on this.

They aren't devoid of bias at all that's clear from the examples here. It's just a mish mash of other's ideas sort of regenerated seems to me to sort of like I Ching mincemeat or I Ching kebab. They aren't more pure they just get blander as they get more refined 'in the machine'.

Elsewhere in your imagery thread when someone was talking excitedly about making a deck you said

I'm not sure that restricting someone else to holding my imaginal vision of a hexagram as an immutable origin of their readings is something I would want to do. It may not be that different from giving my version of a reading in Shared Readings, so it might be an idea to play with in 2024.

...and yet here you seem to be saying the image generated is somehow more pure. It really isn't. It carries one view of 55.5. I agree it is an arresting image but it doesn't hold all that 55.5 is, no image can even if created by a human. So AI images come into exactly the same problem of 'restricting someone else to an imaginal vision...' . As an aside that's no reason not to make art with the hexagrams, if one says making an image is 'restricting someone else to holding your imaginal vision' then not much art or even writing could arise from studying Yi.

At one stage, I even fed in a short passage from one of Hilary's blogs, and told it to imitate her language style. Yeah, it did look halfway plausible. I'm sure Hilary would read through it and say, I'd never say that! But if you didn't know too much, it sounded quite like her.

I find it worrying and quite horrible really that a specialised author like Hilary or Bradford could be robbed in such a way, it seems very cheap indeed to me. Perhaps I am naive about AI, I'm sure I am, but still it's an awful thought that writers can be robbed this way but luckily it sounds like it doesn't work anyway. I've not yet seen any life infused results from this thing it's all very fortune cookie bland.

I'll surely bother you all again with some new example

I don't have anything against you experimenting but I don't want to help, aid or abet a project engaged in robbing real writers to produce some banal mish mash of a reading, it's not a worthy goal. I mean it doesn't seem a worthy goal anyway. I get that some people are interested in the ideas and the process and it may be an intriguing experiment but as a goal, something to work towards it seems a real backwards step. There's lots of interest here so I don't think you will find this discouraging and also please note I really don't doubt your motives, I know you have genuine intent with this I'm just saying the overall goal with AI interpretations seems unworthy to me

But in the end, I decided it was a silly experiment.

Yes. I mean not that you or Remod are silly, I appreciate there's a genuine interest and so on but what possible good reason is there to get a dead thing to interpret Yi when the whole point of it is our actual engagement with interpreting, that's where the life and vitality is.

AI takes the blood out of the process and far from being 'pure' as Mykey puts it it's actually dead. Well there is a purity in objects they are unbiased and 'neutral' because they are not alive. In any case I don't buy this

So I'm putting in a request for another go. I like the idea that readings by AI are devoid of bias and prejudice. They seem to carry a sense of purity about them.

...it just bypasses the real work, the real interaction between Yi and the querent.

 

[remod]

I was learning how to code: a wonderful challenge for a 60 year old

We're almost the same age: I'll be 59 next year.

I hacked away and drew up some fairly ugly code that could reliably print out the text from Wilhelm's translation.

I didn't get if you are still working on this but if you are and want to share anything I'll be very happy.
I put my I Ching related programming stuff in a GitHub repo. The Javascript version (in the index.html file) just points out to multiple translations (of which Hilary's is one) while the Python one picks up the text (just the basic text, no commentaries), if they can be of any help to you, feel free to use and ask for anything than might be unclear.

But yeah. ChatGPT is a confabulation machine. It's absolutely brilliant at producing writing that looks superficially plausible.

Yes, that's what they do and I find astonishing how good they are compared on how simple the basic mechanism they rely on is simple.
That's what is intriguing me.

 

[remod]

There's lots of interest here so I don't think you will find this discouraging and also please note I really don't doubt your motives, I know you have genuine intent with this I'm just saying the overall goal with AI interpretations seems unworthy to me

Thanks @Trojina , I'm not discouraged but I'm quite sad when I'm misunderstood and I'm happy that this is not the case with you . It happened on other sites, (with my research on alternative methods to cast hexagrams) and that left a bitter taste.

In general, not limited to the I Ching, I believe that the infancy of AI that we are witnessing, is posing us difficult questions. I'm doing other experiments (more related to my job) to get a better understanding of AI in general and how we can use it for the better.

Just as a side note, about embarking in this experiment I asked: "Would the experiment of using AI for the I Ching, help me getting a better understanding of what AI is and how I can get benefits from them?"

I got 51.6 >21 which I found extremely relevant with the sixth line being there to warn me about the perils of going on without seeking other's advice. Which is exactly what I'm doing in this thread

Hex 51 suggested me it was a complex and worthy subject to explore and hex 21 convinced me that enduring with the proper judgement will lead me to gaining a little bit of the knowledge I'm seeking.

 

[IrfanK]

@Trojina, yeah, all that.

I recommend a documentary called AlphaGo, about the Google DeepMind Challenge Match with top-ranked Go player Lee Sedo. I don't know if you've come across the game Go? Sort of absolutely nothing like Chess, to which it is often compared. The Koreans and the Japanese hold it incredibly high esteem, and the top-ranked players are revered national heroes.

It's an incredibly moving film. Absolutely no-one believed that a machine could beat a top player, it was thought to be something too ineffable, a skill more like writing poetry than performing mathematical calculations.

Lee Sedo was deeply ashamed and bewildered to be beaten, he apologized profusely for letting down the people who believed in him. The Korean audiences were aghast, people crying and weeping. Even the people who developed the machine seemed to have mixed feelings about what they had done -- you couldn't help feeling that they were actually relieved to see him beat it -- in one match out of five. He resigned as a professional player shortly thereafter and withdrew from the public eye.

I feel like that about ChatGPT, although it's still not really up to human level -- yet. But close, and getting closer. And I've lived off my writing all my life. It's terrifying. It's the very reason that I'm learning computer code, at this age. I have a friend who is quite involved in machine learning, and I kept on badgering him with questions about it. One day he got a bit exasperated and said, I really can't explain this to you unless you know code. So I decided to learn. I want to understand it, and not just be scared of it.

I'm not sure it works. The more I understand it, the more frightening it becomes.

 

[my_key]

Hi Trojina

They aren't devoid of bias at all that's clear from the examples here. It's just a mish mash of other's ideas sort of regenerated seems to me to sort of like I Ching mincemeat or I Ching kebab. They aren't more pure they just get blander as they get more refined 'in the machine'.

By bias I was thinking more as being an inclination towards prejudice either in favour or against a subject which would present the content of the words as being unfair in some way. While in your eyes they may be a mish-mash but in the examples I struggle to see a bias or an unfairness in the content. The examples give, in my eyes, a reasonable interpretation which may well be viewed by some as bland.

I, for one, will hold judgement on whether the blandness becomes more refined 'in the machine' until I have seen more examples. I have read many a human interpretation over the years which carries the characteristics of bias, mish mash and blandness that you offer your vote of disapproval for in this embryonic I Ching oracle.

Elsewhere in your imagery thread when someone was talking excitedly about making a deck you said


...and yet here you seem to be saying the image generated is somehow more pure. It really isn't. It carries one view of 55.5. I agree it is an arresting image but it doesn't hold all that 55.5 is, no image can even if created by a human. So AI images come into exactly the same problem of 'restricting someone else to an imaginal vision...' . As an aside that's no reason not to make art with the hexagrams, if one says making an image is 'restricting someone else to holding your imaginal vision' then not much art or even writing could arise from studying Yi.

I appreciate you taking the time and effort to locate my words from the other thread and copying them here. Placing them side by side together with my words in this thread it's very clear there is a glitch in the matrix.

Yor diligence has prompted me to reflect on what I was trying to say when I posted

"I like the idea that readings by AI are devoid of bias and prejudice. They seem to carry a sense of purity about them. Again, I'm particularly intrigued to see the picture images generated on this."

On reflection, what I was deeply saying was that "I like the idea that readings being devoid of of my personal bias and prejudice that would detract from the purity of the message that comes from Yi". I wasn't able to reach the insight I now have when I first made the post you have resurrected.

Without 'me' in the reading I can see what the message might mean more clearly.
Any unconscious bias regarding the stereotypes and associations I implicitly hold (because they are there for everyone) would be excluded from the message generated by AI. So, although I have used the words 'pure' and 'purity' in earlier posts my latest truth is that I mean 'independent': clear of any of my personal bias - conscious or unconscious.

Thank you for prompting another important step in both the breadth and depth of my personal growth.

I find it worrying and quite horrible really that a specialised author like Hilary or Bradford could be robbed in such a way, it seems very cheap indeed to me. Perhaps I am naive about AI, I'm sure I am, but still it's an awful thought that writers can be robbed this way but luckily it sounds like it doesn't work anyway. I've not yet seen any life infused results from this thing it's all very fortune cookie bland.

In my dealings with Bradford I always found him to be generous in the sharing of his knowledge and understanding of I Ching. I learnt an awful lot from him as I think many of the people in Clarity did 15 or so years ago.

I wonder if he would think he was being robbed or see what is happening here with his work, which he left to mankind, free gratis, would be viewed by him as cheap.

Hilary too has shown great generosity to me over the years, and again I would be surprised if she saw what was happening here was robbing her of something. I wouldn't want to speak for her though as I do not know her mind.

I see what @remod is doing with all of the texts he uses more as an act of respect and honouring of the people who created them.

AI is only as ethical as the people who write the code. There are 'robbers', particularly in the music industry where voices can be emulated and suddenly Robbie Williams, Mariah Carey or any number of other artists have new songs released that they never had any involvement in.

In the same artistic arena AI had a positive involvement in creating the recently released Beatles record featuring the long dead John Lennon on vocals.

AI takes the blood out of the process and far from being 'pure' as Mykey puts it it's actually dead. Well there is a purity in objects they are unbiased and 'neutral' because they are not alive. In any case I don't buy this

I have modified, as above, my thinking and 'pure' is not how I view it now. I have been racking my brains to think of something more pure than the state of death. It's interesting that from your perspective things cannot be alive unless they contain bias and not be in a state of neutrality i.e. are out of balance.

I am glad to hear that you do not buy this idea that comes from my use of the word pure. I used the word erroneously. Sorry to have sent you off down a track following a word that slipped through my lack of understanding of what I really meant.

...it just bypasses the real work, the real interaction between Yi and the querent.

I do agree, in general, with this comment. Doing the real work should never be bypassed. The key to any act of divination is the post cast ' fixing the omen' activity. This is not just the interpretation, but relates more to the reflective, contemplative stages that follow the interpretation and that lead to embracing the inherent wisdom in the oracle. Once the wisdom is seen, further words, deeds or acts can be despatched from deep inside to manifest the change.

I'm not totally sure that getting someone else to make the cast and interpret the reading is totally bypassing the real work. The real work, perhaps, comes in recognising the wisdom and then ensuring that the omen becomes manifest.

For reasons that leave me with a bit of uncertainty about using AI for new readings I have only offered questions that I have already done the 'real work', or at least some of it. I would not offer a question, and certainly not questions that I hold as important, to AI for an initial cast or interpretation.

The questions I have offered are older ones where the comparisons of conscious or unconscious bias can be more easily uncovered. Perhaps, I see the AI as a mirror that will reflect back to me aspects of my work that have been created from a place of imbalance.

Again, I'm not expecting you to agree with anything I have said here. I felt it was appropriate to share dialogue on matters that have not been broached before in this thread.

 

[remod]

A technical update: just modified Yishi to generate lines using the yarrow stalks (16 marbles) probabilities rather than the three coins.
This implies that while the overall chance to get a moving line is 1/4 (as for the three coins method), the chance for a yin line to change is less than the chance for a yang line to change.

 

[my_key]

I'm not sure it works. The more I understand it, the more frightening it becomes.

AI is a two edged sword. It will be, and is already, a great servant for mankind. It will make a terrible and terrifying master though.

I have seen a number of documentaries that show how many big organisations, e.g Apple, Google are using AI in unethical ways that were seen in the documentary as at best manipulative and at worse coercive or abusive. Worse still is the military applications race, (my AI is bigger and better than your AI) and the devastation that wreaks. World powers that are still fearful of the opposition - it's like a modern day version of the arms race. And, perhaps, even worst of all are the countries that are using it for covert facial recognition activities that the only long term objective would seem to be control and the restriction of freedoms of their citizens.

The 'good' or 'bad' in any AI has nothing to do with the AI machine itself, but stems completely from the the goodness or badness in the intent of the programmer that writes the code and what that code actually directs the AI to do.

I don't see anything even slightly whiffing of badness in the experiments that @remod is reporting on here. Generously offering to share his work freely, gives me a warmer feeling than anything I have for either Apple or Google. So I'll continue to support what he is doing in the ways that I can. I fully understand. though, the reasons you have explained for not wanting to support this venture.

 

[my_key]

A technical update: just modified Yishi to generate lines using the yarrow stalks (16 marbles) probabilities rather than the three coins.
This implies that while the overall chance to get a moving line is 1/4 (as for the three coins method), the chance for a yin line to change is less than the chance for a yang line to change.

That's good news. I like the marbles probability emulating closely the yarrow.
If you are looking for a question to test it out on post #63 is still waiting patiently.

 

[Trojina]

In my dealings with Bradford I always found him to be generous in the sharing of his knowledge and understanding of I Ching. I learnt an awful lot from him as I think many of the people in Clarity did 15 or so years ago.

I wonder if he would think he was being robbed or see what is happening here with his work, which he left to mankind, free gratis, would be viewed by him as cheap.

Hilary too has shown great generosity to me over the years, and again I would be surprised if she saw what was happening here was robbing her of something. I wouldn't want to speak for her though as I do not know her mind.

You are responding to a comment taken out of context. When I wrote this part you quoted and responded to I was replying to a particular comment Irfan was making about using AI to copy Hilary's actual style. It was not about simply using translations.

Irfan said

Yeah and the next idea was to see whether ChatGPT could do a half-way convincing interpretation, applied to a real-life question. And the results looked good, at least if you didn't look that carefully. It all sounded plausible enough. I didn't like the style of the language much, so I experimented with that. At one stage, I even fed in a short passage from one of Hilary's blogs, and told it to imitate her language style. Yeah, it did look halfway plausible. I'm sure Hilary would read through it and say, I'd never say that! But if you didn't know too much, it sounded quite like her.

and I responded directly to him

I find it worrying and quite horrible really that a specialised author like Hilary or Bradford could be robbed in such a way, it seems very cheap indeed to me. Perhaps I am naive about AI, I'm sure I am, but still it's an awful thought that writers can be robbed this way but luckily it sounds like it doesn't work anyway. I've not yet seen any life infused results from this thing it's all very fortune cookie bland

So you can see this was not a general comment that there's an issue using an author's translation if credited but Irfan experimented with copying her actual personal style of writing and so potentially that could become plagiarism and a misrepresentation of her work. Fortunately anyone who knows how Hilary writes would likely be able to tell the difference between her and a non human. To be clear I am not saying Irfan did anything wrong, he was experimenting, just that it is worrying overall that there's so much energy going into taking away from a person in order to mimic them. So there is a difference between using someone's translation and copying their whole style which would be plagiarism in my view, misrepresentation and theft of the authors very soul really. Happily it's not there yet and is still producing leaden 'readings'.

Hilary too has shown great generosity to me over the years, and again I would be surprised if she saw what was happening here was robbing her of something. I wouldn't want to speak for her though as I do not know her mind.

Likewise and to me, she has been generous to everyone. But what I said has nothing at all to do with how generous she is I was specifically talking about her very style being mimicked. It is not and cannot ever be a good thing to try to lift someone's entire style and reproduce it. She may not care, no, but in general it doesn't seem a good direction.

I am aware there's a lot to it a lot I don't know about as I have said. I was only saying how it struck me when Irfan described using Hilary's style as an experiment.

 

[remod]

That's good news. I like the marbles probability emulating closely the yarrow.
If you are looking for a question to test it out on post #63 is still waiting patiently.

Sorry I missed that!
Would you prefer a test for an entirely new reading (using marbles probabilities) or the result of the hexagram you already got?

I believe that one is too open ended for the AI to get the intention. What you really want to know is what relationship there is between the too, right?

Because, when I asked that, I got this answer which does not involve directly querying to the I Ching.

I went ahead by asking a new reading, but I thought it was still not focused enough and engaged in a short discussion about it. You can find the chat here.

All in all, I think I prompted Yishi to be more inclined at giving guidance rather than discussing about general topics. I slightly modify the prompt but it doesn't seem it changed his attitude much.

 

[my_key]

Sorry I missed that!
Would you prefer a test for an entirely new reading (using marbles probabilities) or the result of the hexagram you already got?

I believe that one is too open ended for the AI to get the intention. What you really want to know is what relationship there is between the too, right?

Because, when I asked that, I got this answer which does not involve directly querying to the Iching.

Hi remod

I'd really like, in the first place, for the reading I made to be put through the two processes with added pictures. That will help me greatly in my own little experiment.

In addition, there is a curiosity in me that would like the question that I asked in the manner I asked it put through the AI as a virgin question. Just to see what your yarrow probability makes of it. There is, for me, something alluring about what might be said to a 'machine' devoid of heart and soul by Yi. If you have to form a more explicit question then what you say about the core of my inquiry is correct.

I believe that response I get is formed as a consequence of my heart and soul relationship with Yi. So I'm primarily interested in having the AI, heart and soulless, I Ching in the loop.

As a separate exercise what AI has to say about Entropy and Change in general would make an interesting side dish.

Over to you.

 

[my_key]

I am aware there's a lot to it a lot I don't know about as I have said. I was only saying how it struck me when Irfan described using Hilary's style as an experiment.

Hi Trojina
I think that there is a lot that we don't know about an awful lot of things.
Thanks for sharing your perspectives and clarifying where you saw my words as being off piste.

 

[remod]

Hi remod

I'd really like, in the first place, for the reading I made to be put through the two processes with added pictures. That will help me greatly in my own little experiment.

I guess they both missed the point on how to answer.
Here's Yishi (based on Wilhelm translation) and here's Huiyi (based on Hatcher translation) . They both tried to provide advice and I can see why: there is little or no indication in the prompt to allow them to understand the intent.

Note that to minimize hallucinations, I explicitly indicated that the interpretation is to be restricted to the material provided (the question and the hexagram text). I believe there's no room for them to also "guess the question".

This was you were trying to check, right?

In addition, there is a curiosity in me that would like the question that I asked in the manner I asked it put through the AI as a virgin question. Just to see what your yarrow probability makes of it.

Here they are, Yishi and Huiyi still provided a quite unsatisfactory reading (I've made so they interpreted the same response, instead of generating two different ones).

 

[remod]

This latest request took me back to what I was discussing in the other thread: the importance, for me, to formulate the full question explicitly. I'm surely unable to interpret the answer to a question as open ended as "Entropy and Changing lines". I would end up changing the meaning of the question depending on the answer.

If nothing else, I'm getting more clarity on the fact that my approach to the I Ching is to separate the times I "think, learn or discuss about the I Ching" and the times I "ask the I Ching for guidance".

Relating Entropy and Changing lines is an interesting topic but i would have never thought of casting an hexagram to see what would come out. I would have rather searched for reference, suggestions and similar and used to speculate about it.

Maybe, I could have asked a question like "what are the key aspects I should consider when trying to relate the two concepts of Entropy and Cahnging Lines?".

 

[IrfanK]

AI is a two edged sword. It will be, and is already, a great servant for mankind. It will make a terrible and terrifying master though.

I have seen a number of documentaries that show how many big organisations, e.g Apple, Google are using AI in unethical ways that were seen in the documentary as at best manipulative and at worse coercive or abusive. Worse still is the military applications race, (my AI is bigger and better than your AI) and the devastation that wreaks. World powers that are still fearful of the opposition - it's like a modern day version of the arms race. And, perhaps, even worst of all are the countries that are using it for covert facial recognition activities that the only long term objective would seem to be control and the restriction of freedoms of their citizens.

The 'good' or 'bad' in any AI has nothing to do with the AI machine itself, but stems completely from the the goodness or badness in the intent of the programmer that writes the code and what that code actually directs the AI to do.

I don't see anything even slightly whiffing of badness in the experiments that @remod is reporting on here. Generously offering to share his work freely, gives me a warmer feeling than anything I have for either Apple or Google. So I'll continue to support what he is doing in the ways that I can. I fully understand. though, the reasons you have explained for not wanting to support this venture.

Hmmm. Yes, you often hear people solemnly intone that technology is neutral and that it depends on the use to which it is put. It all depends on what people choose to do with it. But I don't really believe that anymore. I think technology creates the power to do things, and that power changes the consciousness of people who hold it.

The most obvious example is a gun. Of course, you could use it for good purposes, as a tool in an agricultural endeavor, for target practice and sport, and for self-defense. But the very fact that you can pick up this tool and, with a minimum amount of effort and fuss, end another human's life, without the ick-factor involved in using a knife or machete, seems to mean that people are more likely to kill each other.

So it is with information technology. The very fact that you can now monitor and watch people, track them through their phones, watch them through CCTV and feed their faces into a database, means that governments WILL do it. In the end, governments will believe that it would be irresponsible for them NOT to do it. After all, all the OTHER governments are doing it.

Well, perhaps in some fantasy dreamed up by Socrates, in a Republic ruled by enlightened philosopher kings, people might use artificial intelligence exclusively to promote happiness, mutual prosperity, the sharing of knowledge.

But back here on Planet Earth, I don't think it's going to unroll that way.

 

[remod]

Well, perhaps in some fantasy dreamed up by Socrates, in a Republic ruled by enlightened philosopher kings, people might use artificial intelligence exclusively to promote happiness, mutual prosperity, the sharing of knowledge.

But back here on Planet Earth, I don't think it's going to unroll that way.

I do agree 100%. AI will affect our lives more and more, not necessarily in a positive way. There will be people whose job will be obsolete or too expensive if left to humans and there is no reason to believe that they will find an alternative in the fields opened by the AI.
Looking back we observed that new technologies always offered, in the long run, better opportunities, but I'm not sure this time will be the same. I always say that "AI raises the bar for us", the unanswered question is how quickly (and to what extent) we'll be able to adapt to this new bar. Both as individual and as society.

The point is not if AI will get "human characteristics" (like our "soul" or "heart"), the point is how "good enough" it will be able to approximate those in our eyes. There is little consolation in saying: "yes, it sounds like a human, it acts like a human, but in the end it is not". In the end if it will smell like a rose it will be considered a rose.

Let me give some examples of areas that triggered my thoughts (no claim I'm right, this is what strikes me the most)
Music. Nowadays most of the "commercial music" has been dumbed down incredibly: Simpler and simpler harmonies, flatter melodies, non-existent lyrics, formulaic songs, ... This is a field that will soon be impacted but not because we'll get back dead singer voices or will fake pop idols, but just because the current state of the music is such that an AI can easily create something that any bedroom producer can (and does) create with some audio plugin.
And this is a good thing! This might push true artists (those who really have the genius to create new music) to emerge while music majors look for ways to do more money. If not, we'll be inundated of AI generated music and AI will kill music.

Exactly as AI killed chess, at least for me. For long time the phrase "computers will never be able to beat men at chess" was considered an indisputable truth. And, in a sense, it still is: a computer can't compute the complete tree of a chess game. The point is that they don't need to, they can just play "well enough" to beat any human today.
And what is interesting is that AlphaZero, one of the two current top chess-playing, was not trained on human played games like Stockfish (the other AI top player). It was given the basic rules and was instructed to play billions of games against itself until its "style" emerged.
Today, world class chess players use AI to study and improve their chess skills. Anyone wanting to seriously study and improve their chess playing ability *must* train using AI.
It's not bad in itself, but, to me, it made the game less interesting. Before AI I could pretend that my ingenuity and skill could raise my playing level higher and higher, with no limit; today there's a cap. It's like if I wanted to outrun a car just by running. Pointless. Chess, that was considered one of the human-only activities, has now entered the realm of "mechanical activities".
The AlphaGo you mention is even more a testimony of this since Go has always been considered more complex to analyze than chess.

I could give other examples but I doubt there's anyone still reading this post ...

Just as a (sort of) conclusion, I believe that AI is pushing more and more activities that were traditional "human only" to the "mechanical activities" area. They don't need to be self-consciuous, they don't need to have a heart or feelings. They jsut need to do those activities "well enough" so that we humans will be satisfied by their outcome.

Talking about this experiment with I Ching, I wanted to check how good (or bad) an AI-generated reading could be. We all noticed the flaws those reading have and will surely spot anyone that will try to pass one those are their own human-created reading.

Much like it happened with the images, after more than one year of being exposed to Midjourney and Dall-e generated images, we can easily identify AI generated images. Unless, and this is the key point, some clever human spent time in prompting and retouching to create somethinng unique using AI as one of the tools at their disposal.

I want to understand, and I want to be able to take advantage of AI so that what I create (whatever it is) will be better than what AI could create alone. I want AI to be one of the tools at my disposal, that's why I'm spending so much time on this topic. By sharing, I hope others will be motivated to explore this at least a little.

I will never be able to stop AI, I don't believe anyone, at this point, will be able to do it. I believe that we should understand AI at a level that will allow us to not be in the position of passively accept whatever governments and major conglomerates will try to push on us.

And, yes, this could be utopic as well ...

P.S. Now that I re-read this, I think I might expand it and publish as a blog post

 

[my_key]

I guess they both missed the point on how to answer.
Here's Yishi (based on Wilhelm translation) and here's Huiyi (based on Hatcher translation) . They both tried to provide advice and I can see why: there is little or no indication in the prompt to allow them to understand the intent.

It's clear the question has to worded to evoke a framework in which the answer can be constructed.
The vagueness of my question was a clear choice for me giving Yi a wide scope for reply. My unspoken intent and all the background information that was bubbling around inside provided me with that insider knowledge to make meaning of the oracle based in the context of the questioning.

Note that to minimize hallucinations, I explicitly indicated that the interpretation is to be restricted to the material provided (the question and the hexagram text). I believe there's no room for them to also "guess the question".

This was you were trying to check, right?

Your two fellas have none of my insider information and it shows in the responses generated. I was really looking to see what would come out of them rather than have any fixed expectations.

Here they are, Yishi and Huiyi still provided a quite unsatisfactory reading (I've made so they interpreted the same response, instead of generating two different ones).

Perhaps endorsing the rubbish in rubbish out view. Questions need to be more helpful than the way I posed it for myself in order for AI to construct something meaningful for the human who created the question.

So rather than asking the vague “Entropy and Changing Lines?“ the more detailed question that was lurking behind this was

"What is the relationship between Entropy and Changing Lines as seen in the context of change that is outlined through I Ching divinations?"

37: 3.4 <> 25

A bit wordier, and I'm not sure if I have captured it as succinctly. I was perhaps entering into inquiry on a meta-physical level. I would like to see what the AI produces in respect of this meatier question. Images too, will have better foundations for their creation.

 

[remod]

"37.3.4 What is the relationship between Entropy and Changing Lines as seen in the context of change that is outlined through I Ching divinations?"

Here is what Yishi and Huiyi got from your question:

Yishi (full reading):

Summary​

The relationship between entropy and changing lines in the I Ching is embodied in the need for balance between structure and flexibility. In Hexagram 37, the family (clan) represents a system where nurturing and perseverance maintain order, akin to the balancing force in entropy that prevents systems from descending into chaos. The changing lines illustrate how both rigidity and laxness can disrupt this balance. Moving to Hexagram 25, the emphasis on innocence (purity of intention) further suggests that true and natural changes align with the fundamental principles of a system, much like how entropy governs the natural order of physical systems. In essence, the I Ching's perspective on change mirrors the principles of entropy: change is constant and must be balanced by forces that maintain harmony and prevent chaos.

The image above represents the themes of entropy, change, balance, and the I Ching as described in your query. It visually embodies the concepts of unity, the spread of influence, the role of a nurturing figure, and the dynamic yet controlled nature of change, all set against a backdrop that symbolizes the balance between structure and flexibility. This artistic representation aims to capture the essence of the I Ching's perspective on change, reflecting the principles of entropy in a harmonious yet dynamic composition.

Huiyi (full reading):

Summary​

The relationship between entropy and changing lines in the context of I Ching divinations, as seen through your question, reflects the dynamic interplay between disorder and order, chaos and harmony. The changing lines in Hexagram 37 indicate that while entropy (in the form of conflict or harshness) can disrupt the family unit, it also offers opportunities for realignment and growth. The shift to Hexagram 25 highlights the importance of authenticity and integrity as stabilizing forces against entropy. In the broader context of change, the I Ching suggests that embracing change, even when it brings temporary disorder, can lead to greater harmony and fulfillment if approached with persistence, honesty, and integrity. Consider how these principles apply to your current situation, and how you might use moments of entropy as opportunities for positive transformation.

Here is the visual representation capturing the essence of dynamic interplay between disorder and order, chaos and harmony, in the context of the I Ching divination related to your question. The image includes symbolic elements representing the family unit, wind and flame, and various aspects of change, growth, and realignment. This artwork metaphorically illustrates how embracing change, even amidst temporary disorder, can lead to greater harmony and fulfillment.

 

[my_key]

Thanks remod.

The meatier question looks to have created a better point of origin from which YiShi and HuiYi can peddle their wares. These I would say are much more satisfactory than the responses you posted before. Vague, sparsely-formed questions are things that both the AI teams struggle with. The answers do become bland and general without something for the AI to get it's teeth into.

Because these responses didn't quite answer my deeper question, which I did my best to formulate this time, I realise now I missed the mark by a degree or two. I can see that there is a real need with AI to spend time contemplating and then formulating a focussed question so that the specific area that the querent is looking for insights into and around is actually fully included and adequately signposted.

My interpretation of the original question, if you want to do any comparisons is in Exploring Divination Thread that I posted about Entropy. My interpretation comes from a direction that could be traced back to help formulate a better starting question. Perhaps, I still don't know what I was really trying to ask.

I spotted 'akin' in the first reading summary. That's one of my favourite words so that gives an extra bonus point to YiShi. It is this reading, too, that resonates with me more than the reading by HuiYi.

 

[breakmov]

Hello Remod
I am following your experience with great interest.
... I'm currently doing a specialization in "robotics and the IOT of things", this has been a very interesting topic from the practical point of view of finding efficient solutions in very complex contexts....

But back to the topic of "Entropy".....

Would it be possible to leave what your program says about:

"define Entropy in the original sense of the term"

"define Entropy in the sense applied to Logical Information Theory"

"define Entropy applied to the system of 64 hexagrams that represents the Yijing"

Thanks

breakmov

 

[remod]

Hi @breakmov! Actually, open ended questions don't work well.
Here is: "define Entropy in the original sense of the term". You will notice that it tries to force fit the concept of Entropy in the readings but since there is no intent in the question, the interpretation has no elements to elaborate.

A question like this, seems more appropriate to be asked to the general GPT model. Since we are here on Clarity, I guess we are asking something like: "Using the I Ching as a framework for your explanation, define Entropy in the original sense of the term." (click for the answer).

"Prompt engineering" is the new term coined to define the task of creating the right initial sentence that will lead to the outcome one is expecting. A vague, open ended prompt is likely to produce a vague and generic output.

In any case here are the answers to other two questions as you posed them:

"define Entropy in the sense applied to Logical Information Theory"

"define Entropy applied to the system of 64 hexagrams that represents the Yijing" (the same answered by GPT4)

PS. I would have loved to spend more time on robotics and IoT, I still hope I'll have enough time after my retirement (which is still loooong way in the future) to dedicate to new and interesting topics.

 

[breakmov]

Remod...simply phenomenal but terrifying at the same time!

....phenomenal but frightening in the sense that there is currently the ability, as your experience shows, to explore and use knowledge made by machines with very good results (thank you)... accessible to the common citizen.
And this is an immense power accessible to any of us, with all that it also requires of ethical responsibility.

All of this brings me to the topic of entropy and its power to evolve into chaos and disorder also applied to AI, which is still taking very young steps.
... for example, the ability to "see what is happening in our visual thoughts", as this experiment shows....

AI tool generates video from brain activity​

, and the possibility (...a speculation) that, in the near future, there will be a greater ability to influence and control the human masses... .what impact will it have on the foundations of Democracy, from an ethical and practical point of view?
...what will our future be like when, for example, a common computer, which processes in bits, starts to work in qubits?...a capacity to relate and process information that is unimaginable, compared to the bit version computer... .the future stopped being a distant place and became an aspect increasingly closer to "today's present".
What would be the reaction if those who created Yijing, calibrated in their way of seeing their world, could appear now and deal with the current world? Would they be shocked at the path their 64 hexagram system took?...
... would they consider an increase in entropy, chaos and destruction of many of their beliefs, or on the contrary a much richer world of information and possibilities and therefore they would be happy to see that "pure change", with no attribute of good or evil, beneficial or not beneficial, continues to be the basic and constant foundation of something mysterious that we call "the indivisible whole".

I am amazed by your experience, it showed very reasonable responses to the topic of entropy, which applied to yijing seems far from easy to me, and really leaves in question what the world will be like in 10 or 20 years.

breakmov

 

[remod]

[...] and really leaves in question what the world will be like in 10 or 20 years.

Indeed! What is really amazing is how this articulated responses could ever been obtained by such a simple mechanism as a LLM. What will happen when more complex architecture will be used? When quantum computing will cut dramatically the time needed for training?

Fascinating, yet scary.

 

[my_key]

Hi @breakmov! Actually, open ended questions don't work well.
Here is: "define Entropy in the original sense of the term". You will notice that it tries to force fit the concept of Entropy in the readings but since there is no intent in the question, the interpretation has no elements to elaborate.

The wording of the question posed to AI certainly seems to be a vital aspect of the enquiry to get right based on what has been on display in this thread.

A question like this, seems more appropriate to be asked to the general GPT model. Since we are here on Clarity, I guess we are asking something like: "Using the I Ching as a framework for your explanation, define Entropy in the original sense of the term." (click for the answer).

This question certainly evokes a reply that gives a very clear answer. Like breakamov, I'm blown away by how this question has been answered. It just demonstrates the potential in AI that awaits harnessing.

"Prompt engineering" is the new term coined to define the task of creating the right initial sentence that will lead to the outcome one is expecting. A vague, open ended prompt is likely to produce a vague and generic output.

How I purposely chose to word my question on Entropy just amplifies the vagueness of my question being something that AI abhors. 'Prompt engineering' is a new one on me but by crafting the question as you have remod, you have certainly made up for my shortcomings.

In any case here are the answers to other two questions as you posed them:

"define Entropy in the sense applied to Logical Information Theory"

Logical Information Theory is not my strong point but I certainly resonate with this sentence in the summary.

In the context of your question about entropy in Logical Information Theory, the I Ching suggests a view of entropy as both a disruptor and a creator of order.

I would even expand it to see 'entropy as both a disruptor and creator of order' in many other fields of activity that the I Ching comments on, both from a personal enquiry and wider perspectives.

"define Entropy applied to the system of 64 hexagrams that represents the Yijing" (the same answered by GPT4)

This time AI casts its own reading with another clear interpretation of the reading. The questions are different and the casts are different, returning 36 unchanging rather than my cast of 37: 3.4 <> 25. There is no reason why they should be the same but I do find it quite amazing that the responses are thematically close.

Entropy, as applied to the Yijing, can be seen as the undercurrents of change within the stability of the hexagrams. The wisdom lies in understanding and harnessing these subtle forces.

My purpose in asking my question was most definitely focussed in the invisible world in and around which 'the undercurrents of change' swirl. Entropy being viewed by AI as being part of this world with an independent cast and an independent interpretation of that cast is for me reassuring.

It removes, for me, any idea of my unconscious bias being present. I look to have created an interpretation of the response I got based in a sense of balance and with a connection to some level of wisdom. It affirms that my reading was not one manipulated by me to obtain the response that I wanted to hear.

remod / breakmov - thank you both for your masterclass in 'prompt engineering'.

A belated Christmas present that I am most grateful to receive.

 

[remod]

I've slightly changed the approach and created a new AI that is more focused on reflecting back the possibilities that are implicit in the question, rather than putting together an "advice".

This is the result of Jī Yí (机疑) interpreting "Should I invest time and resources in starting a Kickstarter project to start selling my I Ching cards?" answered by the I Ching with hexagram 20 (The contemplation).

For comparison, this is the same as interpreted by Yi Shi. They do not seem to disagree on the core interpretation ("this is not the time, work more on it before thinking again about offering your cards for sale") but their approach is quite different.

While Yi Shi tried to reassure me that "it will be well-received and successful", Jī Yí just pushed me to think about it twice: "Have you fully understood the needs and expectations of your potential backers? Are there aspects of your project that need more refinement or clarity before presenting it to the world?"

Maybe this new approach would work better?


BTW: the question is real, and I got similar answers to previous iterations of this question. The I Ching seems to think very highly of my work on "I Ching cards" and encouraged me to go on but every time reminded me that this is something I do for myself, that I should not seek external "approval" but just offer them to whoever is interested, that I should not expect any personal gain, and so on (just to summarize some of the answers I got in the previous years).
It seems Oracle's opinion on my I Ching cards did not change in 2024

 

[my_key]

Hi Remod
A reading that is less based in advice and more held as a clear summary of the wisdom that is being offered by Yi appeals to my perspectives of what Yi is about. A summary of wisdom is more a 'power with' scenario whereas offering advice, while welcomed by some querents, might be more likely to signpost matters in an erroneous direction. It's more powerful for the querent to interpret the wisdom as they see it. Even if it means they go down a 'wrong turning', it is their own wrong turning and not one foisted on them by some well intentioned advice of another - AI or human.

Rather than charging ahead with all guns blazing, 20 does speak more about observing the patterns that are held and still unfolding in the situation. This is what still appears to be the ablution phase of the project, albeit the latter part of that ritual. Ablution has its roots in alchemical practice aligned with ensuring ‘purification through using liquids’. Reminding you to make sure all the dirty washing is properly laundered, and even ironed to remove the creases, in preparation for your pouring out the offering (libation) for true honouring of spirit and the wider populace.

20 asks you to truly find your way to the centre of things and not get lost in the maze, so that like the early Kings you can provide safety, teaching and guidance all focussing smoothly on a common good.

In such times if your intent is not clear and pure within yourself then the thoughts and actions you produce can be at cross purposes. Any attempt you make to launch, in this state, will be premature and most definitely fall between two stools. Your cards will more than likely be viewed as neither fish nor fowl.

Interestingly, the nuclear of 20 is hexagram 23 which guides you towards a direction aligned with new thinking and the stripping away of old ideas.

As an aside, I wonder how you could better use representations of 'liquid based purification' in the ritual of your cards or in the ritual lead up to the Kickstarter project?