Ways to use AI in astrology. Let’s share prompts and reviews on choosing LLMs.

[Xeno]

Let’s discuss ways to use AI in astrology and how we can get the most benefit from it.

Here’s a sample prompt and its response:

You are an astrologer providing answers in a light literary style. Respond in JSON format.
Data:
{
"sun": {'sign': 'Gemini', 'house': '5'},
"moon": {'sign': 'Virgo', 'house': '8'},
"Asc": "Aquarius"
}
Interpretation Plan:
1. "sun_description": "your_response",
2. "moon_description": "your_response",
3. "Asc_description": "your_response",
4. "synthesis1": "summarize descriptions from ('sun', 'Asc')"

******************************************* ******************************************* *******************************************
And here is a response
{
"sun_description": "With your Sun in Gemini in the 5th house, you are naturally curious and expressive, with a strong desire for creativity and enjoyment. You thrive on social interactions and mental stimulation, often finding joy in playful activities and hobbies. Your intellectual curiosity is a driving force, leading you to seek out new experiences and opportunities for self-expression.",

"moon_description": "Your Moon in Virgo in the 8th house suggests a deep, analytical emotional nature. You have a keen eye for detail and a desire to understand the underlying causes of your emotions. This placement can lead to a strong need for emotional security and a tendency to focus on personal transformation and growth. You might find comfort in helping others and addressing complex issues.",

"Asc_description": "With Aquarius rising, you present yourself as innovative and forward-thinking. You have a unique outlook on life and are often seen as someone who challenges conventions. Your approach to new experiences is open-minded and unconventional, and you value individuality and intellectual freedom.",

"synthesis1": "Combining your Gemini Sun and Aquarius Ascendant, you are a highly adaptable and inventive individual. Your natural curiosity and intellectual prowess blend seamlessly with your innovative and unconventional approach to life. You are likely to be seen as a trendsetter, with a knack for engaging with others in a stimulating and dynamic way. Both placements highlight your enthusiasm for learning and your desire to explore new ideas and experiences."
}

p.s.
I would be happy to receive any comments and personal experiences.

 

[Xeno]

Here’s another prompt, but this one is real and was used in the AiDA astrological diary.

sintez_profile = '''Take into account My Data (astrological data, gender and age). Response must be in strict JSON format. Each key and value must be enclosed in double quotes, including nested objects. The JSON schema should include:
{
"start_profile_synthesis": "Create a focused synthesis of a personal general astrological profile using the following steps: Begin with a clear, specific metaphor that connects key elements of my Data to the individual's overall life path and personal characteristics.",
"step_1_sun_moon": "Analyze the Sun's and Moon's signs and houses, detailing their impact on core identity and emotional nature. Provide 2-3 specific insights.",
"step_2_ASC_MC": "Describe 2-3 specific ways the Ascendant and Midheaven influence personal presentation and life direction. Include potential strengths and challenges.",
"step_3_mercury_venus_mars": "Analyze the sign and house of Mercury, Venus, and Mars. Outline 2-3 specific ways these planets interact to affect communication, values, and drive.",
"step_4_jupiter_saturn": "Analyze the sign and house of Jupiter and Saturn. Describe 2-3 specific ways they interact to affect personal growth, life philosophy, and sense of responsibility.",
"step_5_uranus_neptune": "Examine the sign and house of Uranus and Neptune. Indicate 2-3 specific aspects of how they influence the individual's approach to change and spirituality.",
"step_6_key_aspects": "Describe key aspects between significant planets, using metaphors to show their importance for personal development and life experiences.",
"step_final_profile": "Synthesize all elements into a cohesive personal general profile. Compare the individual's life journey to a specific natural process or universal concept. Ensure the synthesis reflects key aspects of my Data, creating a practical and insightful guide to the individual's overall potential, challenges, and life path."
}
Use ✧ symbols to enclose the overarching metaphor in the first step and a rephrased or different personalized affirmation in the final step. For all other steps, avoid using quotes within the generated text. Replace any double quotes within the text with single quotes to maintain JSON validity. Ensure all content is appropriate. After generating the content, perform a final check to ensure all keys and values are properly enclosed in double quotes and the JSON is valid.'''

 

[juliafox1]

I asked horary questions to chatgpt and got a correct answer couple of time. Idk how the calculation worked but I checked and it was correct.

 

[Xeno]

I asked horary questions to chatgpt and got a correct answer couple of time. Idk how the calculation worked but I checked and it was correct.

“Something like a global resonance” ))

 

[juliafox1]

“Something like a global resonance” ))

ChatGPT said that it's not allowed to reach real-time data such of planets etc. so yeah, it could be explained as ''Something like a global resonance” ) life is funny

 

[Xeno]

ChatGPT said that it's not allowed to reach real-time data such of planets etc. so yeah, it could be explained as ''Something like a global resonance” ) life is funny

Of course, we need to include real data in the prompt, as I did in the example.

{
"sun": {'sign': 'Gemini', 'house': '5'},
"moon": {'sign': 'Virgo', 'house': '8'},
"Asc": "Aquarius"
}
there is no other way at least for now

** as for Horary ))
it's smth else or simply
coincidence

 

[waybread]

One wonders how much bread AI took out of the mouths of professional astrologers to tabulate all of this information from their books, blogs, and so forth.

Basically there is a logic to astrological interpretation, which is why Astrodienst www.astro.com has been selling computer-generated horoscopes by Liz Greene for some years.

 

[Xeno]

One wonders how much bread AI took out of the mouths of professional astrologers to tabulate all of this information from their books, blogs, and so forth.

Basically there is a logic to astrological interpretation, which is why Astrodienst www.astro.com has been selling computer-generated horoscopes by Liz Greene for some years.

I think it’s the opposite… The main goal is to popularize astrology, which will not only increase interest but also the number of people seeking astrologers. Simply because people want to communicate with other people. Yes, it’s that simple.


For example, computers play chess better than humans—but people still continue to play in their own tournaments.

As for Astrodienst www.astro.com it's ok to generate horoscopes .. .. but it's wrong example - it just too primitive

 

[waybread]

Xeno, you have a fascinating topic.

My phrase "taking bread out of the mouths of professional astrologers" was tongue-in-cheek. (Now there's a mouthful!)

AI generally is proving to have both positives and negatives.

We read posts from a lot of very vulnerable people here. Nobody turns to astrology to say, "Hey, my life is really going well. How do you explain that?" There is often major stress and self-doubt when people turn to astrology. If AI can be sensitive to people in distress, then it's probably beneficial. If it can't be, then real harm can be done to vulnerable human beings. Hopefully someone has programmed astrological AI with the codes of ethics of the NCGR and ISAR (astrological organizations,) and to clue into what querents' mean but don't actually come out and say.

It will be interesting to see what happens when the bots start talking to each other.

Astrodienst computer generated horoscopes may be primitive compared with AI's expanded reach, but that's not really the issue. The goodness of a horoscope isn't measured by its technological sophistication, but by how well it speaks to an individual's human condition, and how helpful it is to them.

 

[conspiracy theorist]

I have not used LLMs for astrology, but have used them for other things more related to educational and professional pursuits.

Waybread brought up empathy and being able to handle people in vulnerable positions. I have seen people describe https://claude.ai/ and specifically the "Claude 3 Opus" model as having a tone that is more expressive, empathetic and feminine. You could place those "astrological strictures" in the first section of your prompt, while focusing the second section of your prompt on "providing an empathetic and warm tone to a vulnerable person who has recently experienced a nervous breakdown". You can describe the "voice" to be used as a professional therapist with a specialty in crises counseling. Would be interesting to see what response you get.

P.S. I recommend Claude AI over ChatGPT

 

[Xeno]

Xeno, you have a fascinating topic.

My phrase "taking bread out of the mouths of professional astrologers" was tongue-in-cheek. (Now there's a mouthful!)

AI generally is proving to have both positives and negatives.

We read posts from a lot of very vulnerable people here. Nobody turns to astrology to say, "Hey, my life is really going well. How do you explain that?" There is often major stress and self-doubt when people turn to astrology. If AI can be sensitive to people in distress, then it's probably beneficial. If it can't be, then real harm can be done to vulnerable human beings. Hopefully someone has programmed astrological AI with the codes of ethics of the NCGR and ISAR (astrological organizations,) and to clue into what querents' mean but don't actually come out and say.

It will be interesting to see what happens when the bots start talking to each other.

Astrodienst computer generated horoscopes may be primitive compared with AI's expanded reach, but that's not really the issue. The goodness of a horoscope isn't measured by its technological sophistication, but by how well it speaks to an individual's human condition, and how helpful it is to them.

You are saying the right things )

Right now AI by default is programmed to be friendly and ethical, there are built-in system filters.

I even tried to make it less friendly but not too successfully

Another issue that you correctly brought up is the personal context of the user's life, AI simply doesn't know it

• the only solution I found is to integrate AI into an astrological diary (so that there is a place to take context)

And by the way, this is exactly what I'm doing, my pet project (mobile app) AiDA: Astrodiary .. .. still far from ideal but already gives something
really useful.

 

[waybread]

Just as a comment, Xeno, on your Gemini sun -Virgo moon -- Aquarius ascendant example, this is exactly the kind of material found in astrology "cookbooks." It's OK for a start, but misses out a lot of details. A horoscope with an accurate birth time would literally have a few thousand data bits and bytes. I think AI might actually be better than a person at handling the big data set that a horoscope entails; but a simplistic question may yield only a simplistic answer.

A couple of things I look at a lot in reading a nativity are inter-planetary aspects and house cusp rulers Are any planets domiciled? Then there's a judgement call as to which house system one uses; whether one does traditional, Vedic, or modern astrology; what orbs, and so on. AI can catch up to this, but it won't so long as the question focuses on a cookbook approach to planets-in-signs and planets-in-houses.

So one question I would ask, is are that Gemini sun and Virgo moon squared? Are that Gemini sun and Aquarius ascendant trined? Because the answer will make a huge difference to the person's life. Note also that there's a double Mercury influence, so we'd want to see what Mercury is up to. There is also a double air element, suggesting that ideas are very important to this person.

Also, the 5th house is the house of one's hobbies and leisure activities, but it also rules one's children (if any,) "lite" romance (vs. committed relationships,) and generally wanting a spirit of fun in life. The 8th is the traditional house of death. It doesn't mean literal death in any transparent way, but I have noted that 8th house folks tend to think about death more than most people. The 8th also rules "shared resources" like investments. 8th house people also tend to be very private, in my experience.

Then a lot of what people want to know focuses on their relationships, followed by career, money, and friendships. Newbies often ask for a general chart reading, but oftentimes the querent has a specific agenda. If someone asks, "Why does my mother-in-law hate me?" or "Why doesn't my money work out?" there's a way to approach them astrologically.

I'm not saying AI cannot catch up to this level of complexity. Maybe it already has.

 

[conspiracy theorist]

With sufficient input, you could input any ethical code into and LLM and "train" the AI to utilise said code when you prompt it to undertake a certain action. Doesn't need to be programmed into it, either. All that is required is that said code is typed into or "extracted" from a document and placed into the AI. It's not a difficult process.

 

[waybread]

This is all very intriguing.

I have a challenge/suggestion for you. Put up a horoscope with a legitimate question on it. Not a "stump the astrologer" question, but one where the person has something riding on it, and is willing to participate. Then see which AI system -- alongside any astrologers who care to participate -- does the best job according to the chart native.

 

[Xeno]

With sufficient input, you could input any ethical code into and LLM and "train" the AI to utilise said code when you prompt it to undertake a certain action. Doesn't need to be programmed into it, either. All that is required is that said code is typed into or "extracted" from a document and placed into the AI. It's not a difficult process.

Yes ... ... that's easy for ML programmers who know how to train their own LLMs (I'm just learning about this myself).

As for publicly available AIs, they are all filtered or censored - by ethical norms.

 

[Xeno]

Just as a comment, Xeno, on your Gemini sun -Virgo moon -- Aquarius ascendant example, this is exactly the kind of material found in astrology "cookbooks." It's OK for a start, but misses out a lot of details. A horoscope with an accurate birth time would literally have a few thousand data bits and bytes. I think AI might actually be better than a person at handling the big data set that a horoscope entails; but a simplistic question may yield only a simplistic answer.

A couple of things I look at a lot in reading a nativity are inter-planetary aspects and house cusp rulers Are any planets domiciled? Then there's a judgement call as to which house system one uses; whether one does traditional, Vedic, or modern astrology; what orbs, and so on. AI can catch up to this, but it won't so long as the question focuses on a cookbook approach to planets-in-signs and planets-in-houses.

So one question I would ask, is are that Gemini sun and Virgo moon squared? Are that Gemini sun and Aquarius ascendant trined? Because the answer will make a huge difference to the person's life. Note also that there's a double Mercury influence, so we'd want to see what Mercury is up to. There is also a double air element, suggesting that ideas are very important to this person.

Also, the 5th house is the house of one's hobbies and leisure activities, but it also rules one's children (if any,) "lite" romance (vs. committed relationships,) and generally wanting a spirit of fun in life. The 8th is the traditional house of death. It doesn't mean literal death in any transparent way, but I have noted that 8th house folks tend to think about death more than most people. The 8th also rules "shared resources" like investments. 8th house people also tend to be very private, in my experience.

Then a lot of what people want to know focuses on their relationships, followed by career, money, and friendships. Newbies often ask for a general chart reading, but oftentimes the querent has a specific agenda. If someone asks, "Why does my mother-in-law hate me?" or "Why doesn't my money work out?" there's a way to approach them astrologically.

I'm not saying AI cannot catch up to this level of complexity. Maybe it already has.

Actually, AI chooses the answer most frequently found in the numerous sources it has read.

And to adjust the answers - you need to theoretically add personal hints to the program settings.

Something like specifying that the 5th house also indicates children ...

By the way, if children are added to the diary, the problem may be solved by itself without any hints ))

***
As for quiz with choosing a profession, it might also be possible depending on the context in the diary.

• If you provide data on what school and university the person graduated from.
• Gender, age, country of residence (cultural context).
• Perhaps the professions of parents and relatives.
• Their hobbies and other interests.

And then AI will be able to give the most likely scenario.

 

[waybread]

Merven, this thread is on an astrology board. What do you see as the implications of your AI insights for astrology?

 

[Xeno]

AI is a very fascinating subject.

Will AI really come into it's own when sentience emerges? If it EVER does. Will having sentience give it the good bedside manners that skilful practitioners of astrology are able to provide? Technology is sky-rocketing ahead of us.

Ethnics are a HUGE concern for me when it comes to AI. I am thinking of neural nets that rely on the back-propagation method of feeding errors back into the network to make adjustments for more accurate results.

I've had an interest in AI since my youth but haven't been keeping myself updated in recent years. Not sure if things have massively improved besides processing power, speed and storage capacity. The problem for me is the blackbox part of the training process. We know what the inputs are, because they are the training data that's fed in. The outputs are known and are fed back into the system as inputs to make continuous iterative adjustments. The hidden function of the inner layers of the network (the blackbox) is the concern for me. If something goes disastrously wrong and we have no idea how the blackbox part operates, then, how can we fix it via back-engineering a solution. If we remain clueless about its inner workings then we're stuck. Do we just reset and reload and cross our fingers hope for the best?

input (training data) ---> | blackbox | ---> output (processed training data) ----> System reaches an optimum accuracy (training halted)
/|\ ...................................................................................................... |
|_________________ FEED BACK LOOP ________________|

LINK: How Does Back-propagation in a Neural Network Work?

I remember a research article in the New Scientist magazine back in the mid 1990's. It's a long time ago, so my recall might be off.

A research group was creating a HARDWARE neural network circuit based on connectable gated FETs (Field Effect Transistors). The beauty of using the gated FETs was that you could selectively isolate or connect the gated FETs , programmatically via a computer interface. In a similar vein to how our brain neurons with its axons and synaptic, link-up or severe ties to each other when training for particular tasks (not quite the same but you get the gist).

The whole circuit was linked to a computer running an algorithm that would continuously build new network circuits (by connecting / disconnecting the gFETs creating different circuit configurations) and using the back-propagation method to select the optimum circuit.

Here's the interesting bit!

When the researcher's removed the transistors that were isolated (because circuit wasn't using them) the network didn't work as expected. When they re-inserted the isolated transistors the circuit worked as intended. They we're flummoxed. Eventually, they discovered that the capacitance of the isolated transistors was being factored into the whole process by the algorithm causing an interference hampering the function of the net. In the same way that the old analogue televisions experienced interference due to capacitance. The white noise pixelation on the cathode ray screen that occurred any time you stood directly in front of the portable ariel that usually sat on top of it.

My point is could their be similar things happening in the SOFTWARE neural nets of today that we are completely blind to. The problem only discoverable as they happen. How can you provide for contingencies against things like this? You can't! Until things go CAPUT!

TRANSHUMANISM replaces AI and makes it redundant?
Could the futurists have an answer with their take on TRANSHUMANISM. My thinking here is that will cybernetics nullify the need for sentient AI. When humans augmented with advanced brain interfacing technology out compete both humans and AI. The best of both worlds? Can you imagine going to have your birth chart read by a cyborg astrologer (kind of gives me blade runner vibes for some reason). Kind of already kick-started with likes of Elon Musk neuralink chip.

ORGANOIDS? Will organic AI develop sentience and offer a different course?
This is a recent development where stem cells (think embryonic cells that haven't differentiated into the different tissue types) are triggered into transforming into desired cell types. Like neurons or particular organ cells (kidney, liver, heart, etc.). There was an article I read about in one of the science journals (don't remember which now) that used this process to create clumps of neuron balls (Organoids). To create a simple brain (not sure you can even call it a brain) that they interfaced with a computer to play ping pong ball game, lol. Amazing!

LINK: Lab-grown brain cells play video game Pong

Too many things happening. Technology is racing ahead of us and no ones thinking about the ethics of it all. Even though it fascinates me to no end.

I envision a consultation with an astrologer as taking place in a virtual room—one where the astrologer can generate predictive charts in real-time, scroll through statistics, see/hear AI prompts, and conduct the session with the guidance of AI, like a "conductor."

I also think that it's unlikely we'll become true cyborgs... I simply can't imagine how we could accelerate the brain's computations. But wearing AR/VR contact lenses that would do all the possible computing is almost achievable already.

 

[waybread]

Merven, thanks for your thoughtful post.

My mind is boggled.

I view AI as a double-edged sword. Every new technology seems to have this effect on society: some benefits some detriments. Television opened new worlds to viewers in the 1950s, but a big impact was that screen time replaced a lot of human interaction, and replaced it with mind-numbing trivia. Kids spent more time with the television than they did with their fathers. Now kids are glued to their cell phones; yet as a parent, it was a relief to have a faster way to reach them. On my cell phone I have information at my finger tips that used to require a trip to the library. But it can be a huge time-waster, as well. I'm not on social media, because to me, the negatives outweigh the positives. Personal GPS systems have cut down on people's ability to navigate and way-find without an electronic aid.

Pluto in Aquarius will probably show us the dark side of AI, as Aquarius and Uranus modernly rule things electronic and futuristic.

I am wondering about various cognitive biases that might affect AI astrology, as these affect astrology now.

For example, you are aware of the "garbage in, garbage out" problem of scientific research. In psychology, there's confirmation bias, where people tend to select new information as true or false based primarily on their prior beliefs. The Barnum effect is about people tending to believe individual personality interpretations, even though these could just as easily apply to millions of people.

Astrological interpretations, sadly, are often dead wrong. Will AI correct these errors, or simply repeat and reinforce them?

We still don't know why astrology works (assuming it does.) Simple physical causal models (like electromagnetism) fall apart in horary astrology. Astrology as a form of divination presupposes a spiritual metaphysics that would exclude rationalists. This is all material for another post, if not another thread, but I think if we could figure that out, AI would have more of a chance of accurate interpretations and predictions.

 

[Xeno]

Hey @Xeno

Your enthusiasm for AI is infectious.

As a thought experiment I tried to imagine a future, 1000 years from now. Where astrology has been scientifically proven to be a legitimate phenomena and becomes widely accepted.

A future where you only need your birth natal chart and a blood sample when applying for a job. No interviews, no shortlisting for final apprasial. Just the previous two. The birth chart is not a 2d drawing but a futuristic 3d hologram that contains every minutiae you can possibly imagine.

I coined the term micro-astrologics. The advanced AI making correlations with every celestial object's position in the galaxy with a persons birth date. The aspects are 3D within this sphere so the information contained is way beyond what we can imagine or be able to store today.

The aspects would be 3d solids. Think Platonic Solids but more complex and millions upon millions overlapping each other. Capturing the potential, a complex web of future probabilistic possibilities that could arise or be diverted by studying this holographic globe . The future technology would appear to be magic by our standards.



I'll stop before I get too excitable, lol.

Astrology in Fiction

I think we don't need to go for such an expansion

I mean our life span is just too short compared to the cycles of far celestial objects.
So we cannot even theoretically grasp their influence... our garden is just a solar system.

As for the time when we will leave our planet and fly to other stars, we can only speculate about this )))

And thank you for your thinking about AI .. .. Although I think that AI will easily replace most astrologers
It just needs a detailed personal context of the person and he will adapt to it and become the best psychotherapist and best friend

Simply saying there is no limit to self-learning for AI - and at some point it will become very difficult for us to distinguish AI from humans