Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Recurso digital |
| Sprog: | engelsk |
| Udgivet: |
Zenodo
2026
|
| Fag: | |
| Online adgang: | https://doi.org/10.5281/zenodo.19360413 |
| Tags: |
Tilføj Tag
Ingen Tags, Vær først til at tagge denne postø!
|
Indholdsfortegnelse:
- <p><strong>Episode summary:</strong> In this episode of My Weird Prompts, Corn and Herman Poppleberry dive into a fascinating prompt from their housemate Daniel about the future of self-discovery. They explore how 2026 technology has moved beyond rigid personality tests into high-dimensional embedding spaces that map our personal philosophies and political stances with surgical precision. From "Socratic Agents" that reflect your logic back to you to tools that analyze your "semantic drift" over years, the duo discusses how AI can provide a vocabulary for the "politically homeless" and identify our true working styles through behavioral data. They also tackle the thorny issue of algorithmic bias and how adversarial prompting can help us sharpen our own thoughts rather than just confirming them. Tune in to learn how AI is evolving from a productivity tool into a profound mirror for the human soul.</p> <h3>Show Notes</h3> <p>In the latest episode of *My Weird Prompts*, recorded against the backdrop of a clear day in Jerusalem, hosts Corn and Herman Poppleberry take on a profound challenge posed by their housemate, Daniel. The question at hand: Can artificial intelligence move beyond simple task automation to help us define who we are? As we move into early 2026, the conversation suggests that AI is no longer just a tool for writing emails, but a sophisticated mirror capable of reflecting our political beliefs, religious views, and personal philosophies.</p> <p>### From Rigid Quizzes to High-Dimensional Spaces Herman begins by contrasting the "old way" of self-discovery—clunky online quizzes from the late nineties—with modern AI capabilities. Traditional personality tests often rely on linear "decision trees," where a specific answer simply adds points to a pre-defined category. Herman argues that this feels reductive and often reflects the inherent bias of the quiz creator.</p> <p>In contrast, modern large language models (LLMs) utilize what Herman calls "high-dimensional embedding spaces." Instead of placing a person in a box, these models map concepts into a mathematical landscape with thousands of dimensions. This allows for a much more nuanced representation of a person's worldview, where every idea has a specific coordinate relative to every other idea. It is the difference between being labeled "liberal" and having your specific brand of justice mapped somewhere between John Rawls and Robert Nozick.</p> <p>### The Rise of Socratic Agents A key highlight of the discussion is the emergence of "Socratic Agents." Corn and Herman discuss projects like the *Belief Graph*, which facilitates dialogue rather than offering multiple-choice questions. These agents don't just ask if you agree with a statement; they ask *why* you feel a certain way about complex issues like land use or universal basic income.</p> <p>By reflecting a user's logic back to them, the AI identifies underlying principles—such as prioritizing individual autonomy over collective security—without necessarily passing judgment. This provides users with what Corn describes as "intellectual bedfellows." For those who feel "politically homeless," AI can provide the vocabulary and historical context to help them understand that their unique blend of beliefs actually belongs to a rich philosophical tradition.</p> <p>### Personal Data Sovereignty and Semantic Drift The conversation takes a technical turn as Herman explains how individuals are now using "personal data sovereignty tools." By running quantized models locally (like Llama 4), users can feed their own journals, emails, and blog posts into an AI to identify recurring themes and values in their lives.</p> <p>One of the most intriguing tools mentioned is *Eudaimonia*, named after the Greek concept of flourishing. This tool tracks "semantic drift"—the way a person's belief system evolves over years. While Corn notes that seeing one's own evolution can be intimidating, Herman argues that it provides an essential understanding of the "why" behind our personal growth, shattering the illusion of a static self.</p> <p>### Confronting the Bias Problem A major concern with using AI as a philosophical mirror is the inherent bias in training data. Herman acknowledges that most models have a baseline "Western liberal bias." However, he introduces a fascinating countermeasure: "Adversarial Socratic Prompting."</p> <p>By instructing an AI to play the role of a devil's advocate from specific schools of thought—such as Marxism, Neoliberalism, or Distributism—users can cross-examine their own beliefs against a panel of "expert witnesses." This transforms the AI from a potentially biased judge into a set of whetstones used to sharpen the user's own independent thoughts.</p> <p>### Redefining Work and Learning Styles The duo also applies this logic to the professional world. Herman dismisses traditional "learning styles" (like being a "visual learner") as scientifically thin. Instead, he points to tools like *FlowState*, which analyze actual behavioral data—keyboard cadence, document types, and even heart rate—to build a profile of when a person is truly at their most creative.</p> <p>Corn reflects on how this can expose the gap between aspiration and reality. Many people believe they are "morning people" because of social expectations, but data might reveal their high-complexity work actually happens in the late afternoon. This "surgical intervention" by AI offers a level of self-knowledge that generic search engine results simply cannot provide.</p> <p>### Conclusion: From Shouting Matches to Seminars Ultimately, Corn and Herman see these tools as a way to lower the temperature of social and political discourse. In a city as complex as Jerusalem, understanding the foundational principles of one's own beliefs can make a person less defensive when encountering disagreement.</p> <p>If AI can help individuals move from a place of "gut feelings" to a clear vocabulary of virtue ethics or utilitarianism, then a political shouting match can be transformed into a comparative philosophy seminar. By using AI to map the landscape of the self, we may find that we are better equipped to navigate the landscape of the world.</p> <p>Listen online: <a href="https://myweirdprompts.com/episode/ai-mapping-personal-philosophy">https://myweirdprompts.com/episode/ai-mapping-personal-philosophy</a></p>