Hey everyone, welcome back to My Weird Prompts. I am Corn, and I am sitting here in our living room in Jerusalem, looking out at a surprisingly clear day. The light is hitting the limestone buildings just right today. And as always, I am joined by my brother and our resident expert on just about everything.
Herman Poppleberry, at your service. It is good to be here, Corn. We have a really fascinating prompt today from our housemate, Daniel. He was actually talking to us about this over breakfast this morning, and it really got me thinking.
Yeah, Daniel has this way of pinpointing the things we all kind of feel but have not quite articulated yet. He sent us this audio prompt about using artificial intelligence not just to write emails or generate images, but to actually define who we are. Political beliefs, religious views, personal philosophy, even our working styles.
It is a profound shift, right? Think back to how we used to do this. You would go to some website that looked like it was designed in nineteen ninety nine, take a twenty question quiz with a bunch of leading questions, and then it would tell you that you are a left leaning libertarian or whatever. And you always had that nagging feeling that the person who wrote the quiz was just trying to push you into a box.
Exactly. It felt like the result was predetermined by the bias of the creator. But Daniel is suggesting that artificial intelligence might actually remove that human judgment and randomness. It is an interesting paradox, using a machine to understand the most deeply human parts of ourselves.
It really is. And specifically, Daniel was asking about what technologies are actually out there right now, in early twenty twenty six, that go beyond just asking Chat G P T a few questions. He wants to know how we can reliably chart our belief systems and understand ourselves better.
So, Herman, let us start there. When we talk about using A I to map out a personal philosophy or a political stance, what is actually happening under the hood that makes it different from those old school quizzes?
Well, the fundamental difference is the move from a rigid decision tree to a high dimensional embedding space. In those old quizzes, if you answer A to question five, you get ten points for liberalism. It is very linear. But modern large language models, and specifically the specialized tools built on top of them, work by mapping concepts into a mathematical space with thousands of dimensions.
Okay, so instead of a box, it is more like a vast landscape where every idea has a specific coordinate relative to every other idea.
Precisely. And what is really cool is that we are seeing the rise of what I call Socratic Agents. Instead of giving you a multiple choice test, these tools engage you in a dialogue. There is a project called the Belief Graph that has been gaining some traction recently. It is built on the foundations of older collective intelligence tools like Polis, but it uses A I to facilitate the conversation. It does not ask you if you agree with a statement. It asks you why you feel a certain way about a specific issue, like, say, land use in Israel or universal basic income.
I have seen that one. It is interesting because as you talk to it, it starts to identify the underlying principles you are using. It might say, hey, I noticed that in your last three answers, you prioritized individual autonomy over collective security. Does that feel like a core value for you?
Right. It is reflecting your own logic back to you. And that is where the lack of human judgment comes in. The A I is not necessarily saying that autonomy is good or bad. It is just identifying the pattern in your own speech. It is like a mirror that has been trained on every philosophical text ever written, so it can tell you, oh, the way you are describing justice sounds a lot like John Rawls's veil of ignorance, or perhaps it aligns more with Nozick's entitlement theory.
That is the intellectual bedfellows part Daniel mentioned. I love that idea. Being able to say, okay, I have these gut feelings, but where do they fit in the history of human thought? It gives you a vocabulary. I think a lot of people feel politically homeless because they do not have the words to describe their specific mix of beliefs.
And when you are politically homeless, you feel isolated. But if an A I can show you that your specific blend of beliefs actually has a rich tradition, or that there are ten thousand other people in your city who think the same way, that changes your relationship to the world. There are tools now like Consensus and various research oriented wrappers for models like Claude four and G P T five that allow you to upload your own writings or even transcripts of your conversations.
Wait, so you could literally feed it your personal journals from the last five years and ask it to summarize your evolving philosophy?
Absolutely. And people are doing it. There is a growing niche of personal data sovereignty tools where you run a small, quantized model like Llama four locally on your own hardware so your data never leaves your house. You feed it your notes, your emails, your old blog posts, and you ask it, what are the recurring themes in my life? What do I seem to value most when I am under stress versus when I am relaxed?
That sounds incredibly powerful for self exploration. But let us push back on the bias thing for a second. Daniel mentioned that A I could remove human judgment, but we know that these models are trained on human data. Is it really possible to have an unbiased philosophical mirror?
That is the million dollar question, Corn. And honestly, the answer is probably no, not entirely. Every model has a baseline. If you use a model trained primarily on Western internet data, it is going to have a Western liberal bias by default. However, the technology has moved toward what we call adversarial socratic prompting and Constitutional A I.
Tell me more about that. That sounds like something you would be excited about.
Guilty as charged. Basically, you can instruct a model to play the role of a devil's advocate from a specific school of thought. So, if you are exploring your beliefs on economics, you can have the A I represent a traditional Marxist perspective, then a Chicago School neoliberal perspective, and then a Distributist perspective. By interacting with these different viewpoints, the bias of the underlying model is mitigated because you are explicitly asking it to simulate different biases.
So instead of one supposedly neutral judge, you have a panel of expert witnesses that you can cross examine.
Exactly. And that is a much more robust way to find your own center. You are not being told what to think; you are being given a set of whetstones to sharpen your own thoughts against.
I want to pivot a bit to the other part of Daniel's prompt, which was about working and learning styles. I feel like this is an area where people often get stuck in these very basic categories. You know, I am a visual learner or I am an introvert. But those feel so reductive.
They are incredibly reductive, and honestly, a lot of the science behind things like learning styles is actually pretty thin. But what A I can do is observe your actual behavior. There are tools now, often integrated into project management software or even just as browser extensions, that analyze how you actually work.
Like, when do you get distracted? What kind of tasks do you finish ahead of schedule?
Right. There is an A I called FlowState that has been making waves. It does not ask you what your working style is. It monitors your keyboard cadence, the types of documents you are opening, and even your heart rate if you have a wearable connected. Over a month, it builds a profile. It might tell you, Corn, you think you are a morning person, but your most creative, high complexity work actually happens between four and six in the afternoon when you are slightly fatigued.
That is fascinating. It is identifying the reality of my behavior rather than my aspiration. I think I am a morning person because that is what successful people are supposed to be, but the data might say otherwise.
And for learning, imagine an A I tutor that has seen every mistake you have made in a coding language or a foreign language over six months. It does not just say you need to practice verbs. It says, you consistently struggle with the subjunctive mood when you are trying to express doubt, but you nail it when you are expressing desire. Let us tailor the next three lessons specifically to that nuance.
It is that specificity that traditional methods lack. Daniel mentioned that search engines are not great for this. And he is right. If I search how do I learn better, I get a million generic articles. But if I ask an A I that has analyzed my last ten hours of study sessions, I get a surgical intervention.
It is the difference between a map of the world and a G P S that knows exactly where you are standing and which way you are facing. And for the personal philosophy side, it is even deeper. I have been looking into a tool called Eudaimonia. It is named after the Greek concept of flourishing. It basically acts as a long term philosophical companion. You check in with it every day, and over years, it maps the shifts in your belief system using a technique called semantic drift analysis.
That sounds like it could be a bit scary, though. Seeing how much you have changed? Or maybe seeing that you have not changed at all?
It can be. But think about the value of that. Most of us have no idea how our beliefs evolve. We have this illusion of a static self. But if an A I can show you, hey, three years ago, you were very firm on this political stance, but after you had that experience with your new job, you started to soften. It helps you understand the why behind your own evolution.
It makes me think about the Jerusalem context we live in. We are surrounded by so many different belief systems, historical layers, and political tensions. I wonder if a tool like this could help people navigate those conversations more effectively. Like, if I know exactly where my beliefs come from and what their weaknesses are, I might be less defensive when talking to someone who disagrees.
That is exactly what the creators of these tools are hoping for. If you have a clear vocabulary for your own beliefs, you do not feel as threatened by someone else's. You can say, oh, I see, you are coming from a utilitarian framework here, whereas I am looking at this through a lens of virtue ethics. Suddenly, it is not an argument about who is right; it is a discussion about different foundational principles.
It turns a shouting match into a comparative philosophy seminar.
Which, let us be honest, is what the world needs a lot more of right now.
Definitely. I want to go back to the intellectual bedfellows idea. Daniel mentioned finding people who think like you. Are there platforms yet that use A I to connect people based on these deep philosophical profiles?
There are, but they are still in the early stages because of privacy concerns. There is a platform called Synapse that is trying to do this. Instead of a traditional social media profile where you list your hobbies, you have an A I generated belief summary. The platform then suggests groups or individuals who have a high degree of semantic overlap with your core values.
Semantic overlap. That is a very Herman Poppleberry way of saying people who think like you.
Haha, I cannot help it, Corn. But think about the implications. Imagine a political party or a community organization that is not built around a single leader or a vague manifesto, but around a cluster of shared, deeply articulated values that have been mapped out by A I. It could lead to much more coherent and effective collective action.
Or it could lead to even more intense echo chambers. If the A I is only connecting me with people who share my ninety ninth percentile niche beliefs, am I ever going to encounter a challenging idea again?
That is the danger. And that is why the most responsible versions of these tools are being built with what they call diversity injections. The A I is programmed to say, here are five people who agree with you on seventy percent of things, but here is one person who agrees with you on the foundations but has reached a completely different conclusion on the application.
Like a constructive friction. I like that. It is not just about comfort; it is about growth.
Exactly. Growth is the goal. Whether it is understanding why you procrastinate on Tuesdays or why you feel a certain way about the role of the state, the A I is a tool for self mastery.
So, for someone like Daniel, or any of our listeners who want to start doing this right now, what is the practical starting point? If they want to go beyond Chat G P T and really dive into this.
I would say the first step is data collection. Start keeping a digital journal if you do not already. Use a tool like Obsidian or Notion, and then use some of the new A I plugins that allow you to query your own notes. There is a great one called Smart Connections for Obsidian that uses local embeddings to show you how your different notes relate to each other.
That is a low barrier to entry. You are just talking to your own thoughts.
Exactly. Then, if you want to get into the political and philosophical side, I would recommend checking out the Belief Graph project or even just using specialized prompts with a high level model like Claude. Instead of asking it what should I believe, ask it, here are three things I believe; can you identify the common philosophical roots and then provide three strong counterarguments from different traditions?
It is about the quality of the question. The A I is only as good as the prompt, as we always say on this show.
Always. And for working styles, I really think people should look into tools that use passive data rather than self reporting. Anything that integrates with your actual workflow and gives you a weekly summary of your focus periods and your friction points is going to be far more accurate than any personality test.
It is that move from aspiration to reality. I think I am a certain way, but the data shows who I actually am. It is a bit like looking in a mirror in the morning. You might not always like what you see, but it is the only way to start getting ready for the day.
That is a great analogy, Corn. The A I is the mirror. It does not create the face; it just reflects it with a level of detail we have never had access to before.
I also think about the religious aspect Daniel mentioned. That is so sensitive. Can an A I really help you understand your faith?
It is happening. There are projects training models specifically on religious texts, not just to provide answers, but to help people engage with the text in a more personal way. Imagine an A I that has been trained on the Talmud, the New Testament, and the Quran, and you can ask it, how does each of these traditions handle the concept of forgiveness? And then you can reflect on which approach resonates most with your own experience.
It is like having a multi faith panel of scholars available twenty four seven to help you navigate your own spiritual journey.
And again, it is about the removal of human judgment. A human priest or rabbi might have their own agenda or their own interpretation they want to push. The A I can just say, here are the five main ways this has been interpreted over the last two thousand years. Which one speaks to you?
It puts the agency back in the hands of the individual. Which I think is what Daniel was really getting at. Using technology to become more self directed, not less.
Precisely. It is about using A I to augment our self awareness. We spend so much of our lives reacting to things without knowing why. These tools give us a chance to pause and say, oh, I am reacting this way because of this underlying belief I did not even know I had.
It is the ultimate tool for the examined life. Socrates would have had a field day with this.
Oh, Socrates would have been the best prompt engineer in history. Can you imagine?
Haha, he definitely would have. He would have broken the model within five minutes by asking it questions it could not answer.
And that would have been the most valuable part of the session.
Exactly. Well, this has been a really deep dive. I feel like we have covered everything from high dimensional embedding spaces to the future of political parties.
It is a lot to take in, but I think the takeaway is that these tools are here, and they are getting better every day. It is no longer just a gimmick. It is a real path toward self understanding.
And if you are listening and you have tried any of these tools, or if you have your own weird prompts about how A I is changing the way we see ourselves, we would love to hear from you. You can find the contact form on our website at myweirdprompts dot com.
And hey, if you have been enjoying the show, we would really appreciate a quick review on your podcast app or on Spotify. It genuinely helps other people find us and join these conversations.
It really does. We see every review and it means a lot to us. This has been My Weird Prompts. I am Corn.
And I am Herman Poppleberry.
Thanks for listening, everyone. We will talk to you next time.
Goodbye!
