11.7.2024

About the election draws--part two . . .

Quick preview: More interpretive notes, some additional background, and a new AI experiment. Plus, the flash version of my big change decision.

---

Hello everyone. I hope you’ve already read yesterday’s “part 1” post. And here’s a link to the longer explanation of my purpose and methods for the AI/Tarot experiments:

10.27.2024

If you haven’t already read it, please do.

---

Wrapping Up the Election Draws

Yesterday’s comments were focused on the first two sets of draws, which can now be seen as referring to things that have already taken place. The third and fourth sets deals with January events, so we can’t yet look at them in the same way.

Here’s the overview again, for convenience:

As I’ve mentioned previously, the fourth set was done in a separate session, which means that the AI bot did not have any knowledge of the previous sets. That came about because when I did the prompts for set 3, I realized they were not in the best order.

As a reminder—I don’t plan the keywords (like “Democrats” or “January”) ahead of time. I just put them in as they occur to me.

Anyway. I realized that “January” and “2025” were much too broad. I don’t have any idea whether/how the keywords actually work, but that’s the mechanism I’ve been using. And obviously a month could belong to any year, and a year has many months.

So in the final draw of that set, I narrowed down to “January 2025.”

But the three cards in that set seemed didn’t seem to form a coherent group—just on the basis of my intuitive grasp at the time. And now that I can look back at the whole sequence, I can see why.

For reference, here’s the big picture view from the original election draw post:

It turns out that set 3 had no cards in common with the previous two sets—except for the Hierophant, which still wasn’t making sense to me as a running theme.

It works out that set 4 has many cards in common with all three preceding sets, but of course at the time, I hadn’t seen that yet. So I decided to explore the post-election period from a different perspective: what kinds of legal ramifications would there be in January of 2025, in relation to the election.

I was thinking both in terms of legal challenges to the election results, and the effect of the election on court cases pending for Trump. So I started with the keyword “Courts”—but the draw began with the Justice card, and that led me to think the bot might have made a logical connection.

I try to choose keywords that won’t trigger logical associations, and when I see something suspicious along those lines, I discard the results and try a different approach.

In this case, I decided to use “accountability” as a keyword—and the results were unusual in a certain way. Gemini typically produces three drafts in response to each draw request, and most of the time they all have the same three cards. Just some minor differences in interpretation.

The “accountability” draw, however, produced one draft that was completely different from the other two. Here’s what that looked like:

I’m not sure how often this happens, but I only seen it twice before.

I completed the set with a draw on the keyword “aftermath”—and this time Gemini only produced one draft. Which is also quite unusual.

At this point I just felt there was something going on I didn’t understand. And I decided to start a fresh session.

I really don’t have notes to give on the courts/accountability/aftermath series. The anomalies may be significant in a way I can’t see yet, but for now, I’ll just leave those cards up to everyone’s own interpretation.

And that takes us back to the “official” set 4:

Which had a lot of cards in common with the previous three sets, as shown by the color code:

I take those commonalities to be meaningful. And I think it will be interesting to revisit all this in January, and see if any of these draws seem to resonate with actual events.

It may be—as suggested by the Wheel of Fortune—that fate is still unsettled at the present time. We are still two months away from 2025, after all . . . .

---

Unwrapping a Related AI Collaboration

I mentioned yesterday that a Tarot friend had sent me a spread they did on October 25, focused around Trump. Here it is:

That accurately represents the cards he drew, and their positions in a traditional Celtic Cross layout—but I revised the position names so that they would be recognizable to Gemini.

My impulse, as soon as I saw the spread, was to try a new kind of AI/Tarot experiment: asking Gemini to interpret a human-created spread.

I think other people have been doing this sort of thing, but I don’t know what the results have been like. For my part, I’m trying to be very rigorous in designing and tracking these explorations.

In this case, I decided to begin with a completely generic, open-ended prompt:

Let's try an experiment. I am providing you with an image that shows a table, with tarot cards in the first column, and in the second column, the position names for a celtic cross spread. I'm wondering what your interpretation of this reading would be.

The response was quite good I thought, in a general way. It ends with this note:

This is a preliminary interpretation based on the provided image. If you have more information about the querent or specific questions, I can refine the reading further.

So I provided additional information:

Assume the querent is an elderly man, in a powerful position. And focus more on the larger situation, rather than the querent’s feelings.

This time the response was even more interesting—and I asked for some comments about interactions among the cards. After that, I added further information about the querent and the situation:

Could you refine those interpretations a little further using the information that the querent's current situation involves a bitter dispute between two factions in a large organization?

The next response was more complex, and tells a story that fits very closely with current events.

Keep in mind that Gemini knows nothing at all about the querent or the situation beyond what is shown in the cards, and what I’ve put in the prompts. The prompts don’t hint at the querent’s identity or specific position—they just provide accurate, objective, impersonal descriptions that I think anyone (fan or foe) would agree with.

There were two more prompts, which observe the same rules. And by the end . . .

This isn’t exactly a cliffhanger, because I have put the entire exchange in a separate post here on EP. It’s too long for this email, and too nuanced to summarize.

But I really hope you will read it. And when you do, please notice that Gemini draws information from the cards, not from me. Also remember that the cards reflect an intention set by the human reader—but Gemini never knows what that was.

I think you will be fascinated not only by how Gemini approaches the process, but also by the almost uncanny alignment with present reality.

Final thought . . . .

So much of what is written about Tarot today focuses on how humans interpret the cards, whether for self-discovery or as a process with others. My goal is just the opposite—understanding how much information, of what kind, is contained in the cards themselves.

---

Breaking News!

This is the end of what I will write about the election or AI in the EP newsletter. But not the end of my writing about those things, both within and without the Tarot context.

Earlier this week (before worldly things took such a sharp turn) I teased a big change. So the new plan is based on prior reflections.

However—it does seem like a good fit for the new reality in which some of us find ourselves. The complete explanation is quite long, and I didn’t want to detract from this post, so I’ll send out the rest tomorrow.

In the meantime, hope you’ll read An AI/Human Tarot Collaboration.

---

Look for me in your Inbox. And thanks for reading! C