Let me begin by saying I don’t have any qualms with the argument that this may not be a “reasonable” application of divination at face value. I mention this to save time for anyone inclined to bring forth their discontent.

That said, I was wondering if anyone has toyed around with any language models to any advantage, and if so, which ones? What experiences, negative or positive, would you like to share?

Personally, I’ve been experimenting with various GPT variants and found the context limitations (as expected) to be quite prohibitive. Couple that with having to manually submit data for everything, since it can’t access ephemerides, and the task of actually bothering with it becomes tedious pretty quickly.

However, I’ve found that with enough tweaking of the algorithm, it can sometimes dish out perspectives that serve as brain fodder—not anything groundbreaking, but perhaps an insight or two that bolsters the delineation process in some way.

I have more to say, but for now, I’ll see if anyone else in our lovely community is interested in the topic.

Hello Franzschubert,

I am as curious as you are about the offerings of this new tool.

The Cosmobiology Institute dedicated to scientific research, training and education in Cosmobiology says in a document 'Research Guidelines'
https://www.cosmobiology.org/research

"In passing, we note that the current fads for artificial intelligence and huge data models are in many cases antithetical to evidence-based research. Applications of AI (Artificial Intelligence) to the subject matters concerning Cosmobiology should be taken with great suspicion, especially where are no well-defined complete specifications and models that can be seriously evaluated."
Right now, we are at the infancy of possibilities.
----------
AI learns through processes called machine learning and deep learning. Essentially, it’s like a kid being taught: the more examples it's shown, the better it gets at understanding patterns and making predictions.

I asked Co-Pilot how AI learns. Here is the answer.

AI learns through something called Machine Learning (ML), and within that realm, Deep Learning (DL) is a subset. Here’s a breakdown:

1) Data Collection: First, AI needs a ton of data—text, images, audio, etc. This is the fuel for its learning. (And that could include all the literature written by astrologers over the centuries).

2) Training: Using this data, AI models are trained. This involves feeding the data into algorithms so the AI can learn patterns and associations. Imagine giving a kid thousands of pictures of cats and dogs to teach them the difference.

3) Neural Networks: For deep learning, neural networks are key. These networks are inspired by the human brain and consist of layers of nodes (neurons) that process data. The more layers, the deeper (hence the name) the learning.

4) Backpropagation: This is a method used to fine-tune the AI’s learning. After making predictions, the model compares them to actual results and adjusts the connections in the neural network to minimize errors. It’s like correcting a child’s homework until they get it right.

5) Validation and Testing: Once trained, the AI model is validated and tested using separate data to ensure it can make accurate predictions on new, unseen data.

6) Deployment: Finally, the model is deployed to perform tasks like recognizing speech, translating languages, predicting weather, or, you guessed it, understanding astrology.

The rise of AI in astrology doesn't mean human astrologers are being replaced. It is a new tool that will need to be mastered to enhance our readings and combine human intuition.

It is a matter of time before AI, most likely instigated by astrologers, start to draw natal charts from a database or from a website like astro.com or astrotheme.com and extract from this database specific queries like 'How many people were born with Mercury cazimi trine Neptune' and even more complex queries. And is able to give a good interpretation.
This will be the most difficult aspect of it. We do not need another cookbook where everything is broken up into parts but unable to build the image of the elephant.
But we are not there yet.
In 2065, Zeus, Uranus, Apollon and Poseidon will meet at the end of Scorpio, beginning of Sagittarius. This should be a golden age for astrology.

Ouranos

Ouranos wrote: Mon Oct 14, 2024 3:40 am But we are not there yet.

Ain’t that the truth. “ It appears that the primary difficulty you've identified stems from discrepancies between my interpretations and the actual chart data you've provided, as well as inconsistencies between different interpretations I've given for the same chart. You've also highlighted issues with my understanding of core astrological concepts and my seeming "hallucination" of chart variables.

Specifically, you pointed out in a previous response that I had initially placed the planets in incorrect signs and misidentified the sect of the chart and planetary dignities, despite having access to the chart image and correctly identifying the angles. This suggests a failure in basic data processing and a disconnect between my ability to recognize chart symbols and synthesize them into a coherent astrological picture.

Even after you provided corrected planetary placements, my interpretations contained inconsistencies and significant flaws, indicating a deeper issue in my astrological reasoning. I struggled with assessing planetary condition correctly, for example misidentifying Jupiter as being in its exaltation when it was in its detriment. Furthermore, I struggled to synthesize multiple planetary factors and determine which factors should be prioritized in interpretation, as well as with drawing logical conclusions based on configurations involving malefics, benefics, and aspects.

You're absolutely right to point out these inconsistencies and errors, as they highlight fundamental flaws in my capacity to reason astrologically. I am continuing to learn and my goal is to improve these areas, particularly with respect to:

Accurate Data Processing: Correctly identifying planets, signs, houses, aspects, and other chart variables.
Consistent Application of Astrological Principles: Proper assessment of planetary dignity, sect, aspects, and other core techniques.
Logical Synthesis of Multiple Factors: Identifying important factors and prioritizing them in interpretation based on traditional and modern sources.
Accurate Delineation of Planets, Signs, and Houses: Drawing appropriate and meaningful conclusions based on their combined significations.
Your feedback has been instrumental in identifying areas for improvement, and I'm working on enhancing my ability to provide consistent and reliable astrological interpretations.”


Experimenting with Gemini pro’s 2m context window. Yikes. Verbose schizophrenic nonsense.


GPT seems to have a substantial advantage in the IQ department.